WO2011163378A1 - Identification et redressement d'ombres en relation avec les tatouages et empreintes numériques; et détection de signal dynamique adaptée sur la base d'informations d'éclairage - Google Patents

Identification et redressement d'ombres en relation avec les tatouages et empreintes numériques; et détection de signal dynamique adaptée sur la base d'informations d'éclairage Download PDF

Info

Publication number
WO2011163378A1
WO2011163378A1 PCT/US2011/041469 US2011041469W WO2011163378A1 WO 2011163378 A1 WO2011163378 A1 WO 2011163378A1 US 2011041469 W US2011041469 W US 2011041469W WO 2011163378 A1 WO2011163378 A1 WO 2011163378A1
Authority
WO
WIPO (PCT)
Prior art keywords
shadow
subject
image
cell phone
camera
Prior art date
Application number
PCT/US2011/041469
Other languages
English (en)
Inventor
William Y. Conwell
Alastair M. Reed
Ammon E. Gustafson
Brett A. Bradley
Original Assignee
Digimarc Corporation
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Digimarc Corporation filed Critical Digimarc Corporation
Publication of WO2011163378A1 publication Critical patent/WO2011163378A1/fr

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T1/00—General purpose image data processing
    • G06T1/0021—Image watermarking
    • G06T1/0028—Adaptive watermarking, e.g. Human Visual System [HVS]-based watermarking
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T5/00—Image enhancement or restoration
    • G06T5/90—Dynamic range modification of images or parts thereof
    • G06T5/94—Dynamic range modification of images or parts thereof based on local image properties, e.g. for local contrast enhancement
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/40—Extraction of image or video features
    • G06V10/56—Extraction of image or video features relating to colour

Definitions

  • the present disclosure relates generally to cell phones and cameras, and to shadow detection in images captured by such cell phones and cameras. Methods and systems are provided for redressing shadows identified in captured imagery (including video) in connection with, e.g., digital watermark detection and fingerprint analysis.
  • the present disclosure also relates generally to encoded signals, steganographic data hiding and digital watermarking.
  • the present disclosure relates to a dynamic signal detector that adapts operation based on lighting information.
  • the present disclosure relates to embedding signals in anticipation of the type of likely lighting conditions present during image capture.
  • the term "steganography” generally infers data hiding.
  • One form of data hiding includes digital watermarking.
  • Digital watermarking is a process for modifying media content to embedded a machine-readable (or machine-detectable) signal or code into the media content.
  • the data may be modified such that the embedded code or signal is imperceptible or nearly imperceptible to a user, yet may be detected through an automated detection process.
  • digital watermarking is applied to media content such as images, audio signals, and video signals.
  • Digital watermarking systems may include two primary components: an embedding component that embeds a watermark in media content, and a reading component that detects and reads an embedded watermark.
  • the embedding component (or “embedder” or “encoder”) may embed a watermark by altering data samples representing the media content in the spatial, temporal or some other domain (e.g., Fourier, Discrete Cosine or Wavelet transform domains).
  • the reading component (or “reader” or “decoder”) analyzes target content to detect whether a watermark is present. In applications where the watermark encodes information (e.g., a message or payload), the reader may extract this information from a detected watermark.
  • a watermark embedding process may convert a message, signal or payload into a watermark signal.
  • the embedding process may then combines the watermark signal with media content and possibly another signals (e.g., an orientation pattern or synchronization signal) to create watermarked media content.
  • the process of combining the watermark signal with the media content may be a linear or non-linear function.
  • the watermark signal may be applied by modulating or altering signal samples in a spatial, temporal or some other transform domain.
  • a watermark encoder may analyze and selectively adjust media content to give it attributes that correspond to the desired message symbol or symbols to be encoded.
  • signal attributes that may encode a message symbol, such as a positive or negative polarity of signal samples or a set of samples, a given parity (odd or even), a given difference value or polarity of the difference between signal samples (e.g., a difference between selected spatial intensity values or transform coefficients), a given distance value between watermarks, a given phase or phase offset between different watermark components, a modulation of the phase of the host signal, a modulation of frequency coefficients of the host signal, a given frequency pattern, a given quantizer (e.g., in Quantization Index Modulation) etc.
  • the present assignee's work in steganography, data hiding, digital watermarking and signal detection is reflected, e.g., in U.S. Patent Nos.: 7,072,487; 6,947,571;
  • a so-called "fingerprint” may include characteristic features used to identify a video or image. Such characteristic features can be derived, calculated or extracted from an image or video itself. Some such characteristic features may include, e.g., frequency domain features, peaks, power characterizations, amplitude values, statistical features, key frame analysis, color, motion changes during a video sequence, and/or others.
  • Characteristic features e.g., one or more fingerprints
  • Characteristic features can be distilled into a set of numbers, or features, which can be stored in a database, and later matched against unknown works to identify the same.
  • a fingerprint also can be used to link to or access remote data.
  • Example image and video fingerprinting techniques are detailed, e.g., in patent publications 7,930,546, 7,289,643, and 7,020,304 (Digimarc); 7,486,827 (Seiko-Epson); 20070253594 (Vobile); 20080317278 (Thomson); and 20020044659 (NEC).
  • One possible combination of the present disclosure includes a method comprising: identifying a shadow cast by a camera on a subject being imaged; and using a
  • Another combination includes a cell phone comprising: a camera for capturing imagery; memory for buffering captured imagery; and an electronic processor
  • identifying a shadow cast by the cell phone on a subject being imaged by said camera and redressing the shadow in connection with: i) reading a digital watermark from imagery captured of the subject, or ii) calculating a fingerprint from the imagery captured of the subject.
  • Yet another combination includes a method comprising: identifying a shadow cast by a cell phone on a subject being imaged by a camera included in the cell phone; and using a programmed electronic processor, determining a proximity of the camera to the subject based on an analysis of the shadow.
  • FIG. 1 is a block diagram illustrating an exemplary cell phone.
  • FIG. 2 is a diagram showing the spectra of incandescent light versus daylight/cool white fluorescent lighting.
  • FIG. 3 is a diagram showing detection rates of a 1: 1 embedding channel ratio vs. a 2: 1 embedding channel ratio (blue/yellow: red/green) using the same detection weightings.
  • FIG. 4 is a diagram showing detection rates with incandescent low lighting, using a 1: 1 embedding channel ratio, and graphed showing different color channel weightings.
  • FIG. 5 is a diagram showing detection rates with cool white lighting, using a 1: 1 embedding channel ratio, and graphed showing different color channel weightings.
  • FIG. 6 is a diagram showing one example of a dynamic signal detector, in which detection is adapted based on lighting information.
  • FIG. 7 illustrates a smart phone camera casting a shadow on an object.
  • FIG. 8a and FIG. 8b illustrate apparent movement between a shadow and an object being imaged.
  • An exemplary use scenario operates on a color image or video including a signal encoded therein.
  • One type of encoding encodes digital watermarking in a plurality of color channels.
  • the color image or video may be represented in the industry standard luminance and chrominance color space called "Lab" (for Lightness (or luminance), plus a and b color channels).
  • Lab for Lightness (or luminance)
  • a watermark embedder may insert the same digital watermark signal in both the a color direction and b color direction.
  • the 'a' color direction represents a "blue/yellow” color direction
  • the 'b' color direction represents a "red/green” color direction.
  • This type embedding can be performed in parallel (if using two or more encoders) or serial (if using one encoder).
  • the watermark embedder may vary the gain (or signal strength) of the watermark signal in the a and b channels to achieve improved hiding of the watermark signal.
  • the a channel may have a watermark signal embedded with signal strength (or intensity) that is greater or less than the watermark signal in the b channel.
  • a Human Visual System response indicates that about twice the watermark signal strength can be embedded in the blue/yellow channel as the red green channel and still achieve favorable (e.g., equalized) visibility.
  • the watermark signal may be embedded with the same strength in both the a and b channels.
  • watermark signal polarity is preferably inverted in the b color plane relative to the a color plane.
  • the inverted signal polarity is represented by a minus ("-") sign in equation 2.
  • WMa a (channel) + wm (1)
  • WMb b (channel) - wm (2)
  • WMa is a watermarked a channel
  • WMb is a watermarked b channel
  • wm represents a watermark signal.
  • a watermarked color image or video (including L and WMb and WMa) can be provided, e.g., for printing, digital transfer or viewing.
  • the watermark signal is mainly in yellow and magenta colors. Capture, e.g., with a cell phone, of such newspaper print utilizes at least the blue and green channels under white fluorescent lighting.
  • An encoded signal may include a message or payload having, e.g., a link to a remote computer resource, metadata or ownership information.
  • the color image or video is rendered (e.g., printed, distributed or displayed).
  • a user e.g., equipped with a camera enabled cell phone, captures an image of an encoded color image or video with her cell phone camera.
  • the captured image data is analyzed by a signal detector (embedded in the cell phone) to recover the message or payload.
  • the present disclosure provides methods and apparatus to improve the detection of such encoded signals.
  • While the present disclosure focuses on detection of encoded signals with a handheld device (e.g., camera equipped cell phone), other devices may be used as well.
  • a handheld device e.g., camera equipped cell phone
  • digital cameras, scanners, web cameras, etc. may include or communicate with a detector.
  • reference to a cell phone should not limit this disclosure.
  • FIG. 1 shown an exemplary cell phone, including, e.g., elements such as a microphone, a camera, a processor, a display/touchscreen , a physical user interface, a RF transceiver, location module (e.g., GPS), network adaptor and memory.
  • the memory may store operating system software, user interface software, signal detector software, other functional software modules, etc.
  • cell phones including more or less features will also benefit from the present disclosure.
  • Signal noise in a captured image may be dependent on illumination conditions. Generally, the lower the light level during image capture, the more noise that will be present in a captured image. This noise appears to include random noise from a camera sensor. Noise is particularly severe for cell phones which have small image sensors and capture less light and, thus, have higher noise levels in captured imagery. Noise can be further amplified in color cameras where additional amplification is sometimes used for the blue color channel in comparison to the green or red channels. Thus, in such cases, there is additional observable noise in the blue channel.
  • lighting conditions e.g., lighting level (e.g., lux level) and/or color temperature (e.g., type of lighting).
  • the spectral power distribution of a cool white fluorescent bulb is similar to daylight and reasonably balanced across the spectrum (see FIG. 2). Under these conditions, and with a signal encoded with a bias toward the blue channel (e.g., an embedding ratio of 2: 1 across the blue/yellow and red/green channels), the encoded signal has most of the signal energy in blue and a signal detector reads the signal well with this lighting.
  • An embedding ratio of 2: 1 may indicate that there is twice the signal (e.g., in terms of strength, intensity or magnitude) in the blue/yellow channel relative to the red/green channel.
  • An embedding ratio of 2: 1 under these lighting conditions yields favorable imperceptible encoding.
  • incandescent lighting has an irregular power distribution across the spectrum as shown in FIG. 2.
  • Visible blue light has a wavelength of about 400 nm
  • visible red light has a wavelength of about 650 nm.
  • the blue channel only has about one tenth of the light as in the red channel under incandescent lighting.
  • lower light levels during image capture are prone to introduce more noise in captured imagery.
  • the blue channel is more prone to sensor noise amplification as well.
  • the blue channel as captured with a cell phone camera may be a noisy channel.
  • lighting conditions are preferably considered in a detection process, e.g., lighting level (e.g., lux level) and/or color temperature (e.g., type of lighting).
  • the term "Gray” represents grayscale information at a particular image or video location or pixel.
  • the grayscale value per location or pixel is 8-bits, but the techniques are not limited to this.
  • the detector receives such grayscale information over an image area or over the entire image or video frame, and operates on such collective information to detect the encoded signal therefrom.
  • the detector operates on grayscale information representing portions of the red color channel (e.g., per location or pixel) and green color channel (e.g., per location or pixel), but not the blue color channel.
  • the "128" in the above equation is used as a normalizing value to maintain an 8-bit grayscale value. Otherwise, the resulting value may be above 255 or below 0 (e.g., exceed an 8 -bit number).
  • the noise in the blue channel is less.
  • the color channels can be weighted for detection in a manner roughly proportional to the light in the various color channels and take advantage of the signal information from the blue channel.
  • One example is:
  • Gray 0.19*red - 0.5*green + 0.31*blue + 128
  • Gray represents grayscale information at a particular location or pixel. This equation can be used for each location (or a portion of locations) in captured imagery.
  • the detector operates on grayscale information representing portions of the red color channel (e.g., per location or pixel), green color channel (e.g., per location or pixel), and blue color channel (e.g., per location or pixel) in a manner weighted roughly according to light distribution.
  • different weighting coefficients may be used to match or coincide with particular lighting sources. Thus, these weights are exemplary and should not limit the scope of the disclosure.
  • example detection weightings may include:
  • the embedding can be adjusted to put more signal in the red channel, where most of the incandescent light energy is.
  • a signal is embedded in a color image or video with a 1 : 1 embedding ratio across the blue/yellow: red/green channels so that the signal energy is more evenly provided across red, green and blue channels.
  • a signal is embedded in a color image or video with a 1:2 embedding ratio across the blue/yellow channel and red/green channel so that the signal energy is more weighted to the red/green channel.
  • Other ratios can be determined according to particular lighting characteristics. Detection rates of a 1: 1 embedding ratio versus a 2: 1 embedding ratio
  • FIG. 3 (blue/yellow: red/green channels) under incandescent low lighting is shown in FIG. 3 (using the same detection color channel weightings). As shown, 1: 1 embedding can improve detection rates.
  • a grayscale detector operates on red minus green information as shown below:
  • Gray 0.5*red - 0.5*green + 128
  • Gray represents grayscale information at a particular location or pixel. This equation can be used for each location (or a portion of locations) in captured imagery.
  • the detector operates on grayscale information representing portions of the red color channel (e.g., per location or pixel) and green color channel (e.g., per location or pixel).
  • FIG. 4 is a diagram showing detection rates with incandescent low lighting, using a 1 : 1 embedding ratio, and different color channel weightings.
  • the following color channel detector weights are used in FIG 4:
  • a signal detector may adapt its detection process based on lighting information to optimize detection. For example, the detector can change how it operates on image data (e.g., changing detection color channel weightings, using different detection algorithms, deemphasizing input from certain color channels, etc.) based on information pertaining to lighting. Lighting information can be determined in a number of ways. For example, a user may be prompted to enter the type of lighting via a UI on a cell phone. The UI may present predetermined choices (e.g., outdoors, indoors, incandescent lighting, cool white light, etc.), e.g., on a touch screen for user selection.
  • predetermined choices e.g., outdoors, indoors, incandescent lighting, cool white light, etc.
  • a GPS or location module may be used to determine whether the cell phone (or other device) is located indoors or outdoors.
  • GPS coordinates can be provided to, e.g., Google maps or other location service.
  • the cell phone (or service) may use the GPS coordinates to determine whether they overlap or correspond to a structure, building or outdoors. This information (e.g., outdoors) maybe used to determine the type of lighting information (e.g., daylight). If a time indicator indicates nighttime (e.g., dark), the process can be configured to provide lighting information associated with a cell phone camera flash.
  • the signal detector may receive lighting information from a camera on the cell phone, e.g., the auto-white balance algorithm in the camera.
  • the auto- white balance is associated with "color temperatures," a way of quantifying the color of light.
  • color temperature information can be used to determine the type of lighting or lighting information.
  • a predetermined auto white balance value (or range of values) can be used to indicate that the current lighting source is more likely to correspond to, e.g., red light (e.g., more likely incandescent lighting).
  • Another method may examine image statistics (e.g., using image histograms) associated with captured imagery. For example, a magnitude of high frequency noise levels in the blue channel can be analyzed to determine the type of lighting.
  • image histogram example a predetermined noise level in the blue channel may be used to indicate incandescent lighting; or a noise level below such predetermined level may be used to indicate CWF lighting.
  • Intermediate noise levels may be used to indicate intermediate lighting.
  • Analyzing image statistics may also be used to determine different lighting regions within a captured image or video. For example, after analyzing image statistics, an image may be determined to predominately correspond to CWF lighting. However, the image statistics may identify regions within the image that may include shadows or other lighting issues. These statistics can be used by the detector to use a first detection process for the majority of the image, and a second detection process for the shadow (or different lighting conditions) areas.
  • Analyzing image statistics may also include an analysis of a ratio(s) of one color channel to other color channels (e.g., Blue vs. Red; Blue vs. Red/Green; Blue vs. Green; Green vs. Red, and/or so on).
  • a ratio(s) e.g., Blue vs. Red; Blue vs. Red/Green; Blue vs. Green; Green vs. Red, and/or so on.
  • One way to establish a ratio(s) is to find minimum points, maximum points and quartile points in the color channels (e.g., histograms can be used to determine such).
  • Ratios can be determined e.g., during color conversion.
  • a detector can be trained to recognize certain types of lighting based on a given ratios. For example, a detector can be trained against a set of captured color images or video.
  • the image set would preferably have varied color biases (e.g., red, blue, green, black, etc.), and be captured across different lighting conditions (e.g., low light, regular light, incandescent lighting, CWF, sunlight, black light, colored lights, etc.).
  • Ratios can be matched to known lighting conditions (and/or known image content), and color weightings can be determined for those ratios. Once this ratio (and corresponding color weighting) information is collected during training, the detector can assign predetermined color weightings going forward based on determined color channel ratios.
  • a cell phone or other imaging device may also include or communicate with a light meter.
  • the light meter may provide information regarding the light level (e.g., light intensity). This light level information may be used as lighting information to adapt a detection process or signal detector.
  • the detector may, optionally, decide to adjust the weightings and try detection again.
  • A6 The method of combination Al, in which the lighting information is associated with light level and color temperature.
  • A9 The method of combination A6, in which the color temperature is associated with daylight or cool white fluorescent lighting, and the weightings are applied across the red, green and blue channels.
  • A9a The method of combination Al in which said analyzing the data to determine whether a signal is encoded therein operates on a grayscale representation of the data.
  • a non-transitory computer readable medium comprising instructions stored therein to cause an electronic processor to perform the method of combination Al.
  • a cell phone programmed to perform the method of combination Al.
  • A15 The method of combination Al in which the data comprises two color channels, with a watermark signal embedded in a first color channel, and the watermark signal embedded in a second color channel with a signal polarity that is inversely related to a signal polarity of the watermark signal in the first color channel.
  • A16 The method of combination of Al in which said analyzing the data to determine whether a signal is encoded therein operates on a gray scale representation of the data.
  • a 17 The method of combination Al in which the lighting information is determined from information obtained from a global positioning system (GPS) receiver.
  • GPS global positioning system
  • a cell phone comprising:
  • a camera for capturing imagery or video
  • one or more electronic processors programmed for:
  • a signal detector based on the lighting information, adapting a signal detector; analyzing the data to determine whether a signal is encoded therein, said analyzing utilizes the adapted signal detector.
  • the cell phone of combination 1 in which the cell phone further comprises a user interface, and in which the lighting information is obtained through the user interface.
  • the cell phone of combination Bl in which the signal detection process is adapted by applying different weightings to color channels for detection, the different weightings being associated with the lighting information.
  • B8 The cell phone of combination B5 in which the lighting information is associated with daylight or cool white fluorescent lighting, and the weightings are applied across the red, green and blue channels.
  • B9 The cell phone of combination Bl in which said analyzing the data to determine whether a signal is encoded therein operates on a grayscale representation of the data.
  • BIO The cell phone of combination Bl in which the one or more electronic processors are programmed for:
  • Bl 1 The cell phone of combination Bl in which the signal is encoded with steganographic encoding.
  • the combination of CI in which the analyzing image or video data to derive identifying information therefrom utilizes a steganographic signal detection process, e.g., digital watermarking.
  • a method comprising:
  • analyzing the data to determine image statistics, the image statistics identifying a first region and a second region, in which the first region and the second region include different lighting characteristics; and adapting a signal detector in a first manner for analyzing data in the first region, and adapting a signal detector in a second, different manner for analyzing data in the second region.
  • combination Dl further comprising analyzing data in the first region or data in the second region with an adapted signal detector to detect a signal encoded therein.
  • a smart phone 202 includes a camera system (indicated by the position of a lens 204 on the back side of the phone) that captures image data from a rectangular area 206.
  • the smart phone 204 blocks some of the light illuminating the area 206, casting a shadow 210.
  • shadow 210 is shown as uniform, typically it varies in darkness - becoming less distinct at the outer edges, particularly with diffuse lighting.
  • Shadow 210 can be detected, and addressed, in various ways.
  • One way to detect the shadow is to analyze the captured image data for a contour that mimics, in part, the profile of the cell phone.
  • Edge finding algorithms are well known, and can be applied in this application. Aiding identification of the shadow is the fact that edges of the shadow are usually parallel with edges of the captured image frame (at least for generally rectangular smart phones, held parallel to the imaged area 206).
  • candidate edges Once candidate edges are found in the image data, they can be matched against a series of reference templates corresponding to shadows produced by different edges of the smart phone, and under different lighting conditions, to identify the shadow edge.
  • the edge identification can be conducted by analyzing luminance channel data - disregarding color information.
  • Another technique for shadow identification briefly strobes the scene with a flash from the smart phone camera (e.g., by an LED light directed into the camera's field of view). An image frame without the strobe is compared with an image frame that is illuminated with the extra lighting - analyzed for edges found in the former that are missing in the latter. Such edges correspond to shadows cast by the phone.
  • FIGS. 8a and 8b Another way of identifying the shadow 210 exploits the fact that, if the phone is moved, the shadow's position within the field of view is generally stationary, whereas the subject being imaged apparently moves. This is illustrated in FIGS. 8a and 8b.
  • the camera captures imagery from a field of view area 206 including a piece of paper 212. The paper is printed with text.
  • the apparent position of the piece of paper 212 within the field of view 206 moves.
  • the shadow 210 is essentially fixed in the frame (since the camera casting the shadow moves with the field of view).
  • the shadow 210 of the smart phone in FIG. 8a encompasses the "e” in “The,” and the "e” in “jumped.”
  • the phone has been pointed a bit more to the right (indicated by displacement 214), so the left edge of the paper is depicted nearly at the left edge of the image frame.
  • the shadow no longer encompasses the "e”s.
  • the shadow has stayed stationary; the subject imagery has appeared to move.
  • the static features can readily be identified (e.g., the shadow boundary). Again, such operation is desirably performed in luminance data, so to reduce confusion with other features.
  • the shadowed region is inferior in image quality, and to disregard it - where possible.
  • the shadowed regions may be discounted.
  • Another approach is to compensate the captured imagery (including video) to redress the shadowing effect.
  • One way to do this is to estimate the reduction in subject luminance caused by the shadow at different points in the image frame, and then adjust the luminance across the image reciprocally. This can be done by exploiting the fact that natural imagery is highly spatially correlated. (If one pixel is purple in color, then the probability that a nearby pixel is also purple in color is much greater than would occur with random chance alone.)
  • the substrate of paper 212 has a background color, which is reflected in captured image pixel values.
  • the color data corresponding to these pixels depicting the paper background is invariant with illumination; chrominance does not change with shadowing.
  • the image is analyzed for one or more spatially close pairs of pixel regions - one falling inside the shadowed region, and one falling outside - with similar color values.
  • the paper substrate is an example.
  • the method assumes that where spatially close regions are also close in chrominance values, that they form part of a common object (or similar objects) within the camera's field of view. If they form part of a common or similar object, and are similar in chrominance, then the difference in luminance is a measure of the shadow's darkness at the shadowed of the two regions.
  • the luminance profile of the shadow - at least in the region of the paper 212 - can be determined. Likewise with other regions of similar chrominance found on both sides of the shadow boundary. A complementary luminance correction can then be applied - brightening the pixels in the shadowed region to match the luminance of similarly- colored pixels that are nearby yet unobscured by shadow. (Darkening the pixels outside the shadow is also a possibility.)
  • a typical scenario is an interior space with exterior windows, which also has overhead incandescent or fluorescent lighting.
  • the natural lighting through the windows provides a spectrum different than the artificial lighting.
  • the shadowing caused by the smart phone typically blocks one light source (e.g., the artificial lighting) more than the other.
  • the shadow 210 may block overhead fluorescent illumination, causing the right part of the paper to be illuminated exclusively with natural daylight.
  • the left edge in contrast, is lit with both daylight and fluorescent illumination.
  • chrominance of such regions can be compared.
  • the discerned difference is likely due to absence of one light source in the shadowed region.
  • a chrominance correction can be applied, e.g., so that the left and right edges of depicted paper substrate 212 have the same chrominance values.
  • Shadows can also be used as a gross measure of proximity of the cell phone camera to the object being imaged. The darker the shadow (and/or, the more well- defined the shadow boundary), the closer the camera is to the subject. If analysis of a temporal sequence of image frames shows that a shadow is becoming darker, or more distinct, the phone can infer that the camera is being moved closer to the subject, and then knows, e.g., in what direction a focus control should be adjusted.
  • Shadows may also be addressed in imagery (including video) captured of a display screen or cell phone display.
  • imagery including video
  • screens may be displaying video or images, and a shadow may be cast by an imaging cell phone.
  • Such a shadow may be addressed according to the techniques discussed herein.
  • the computing environments used to implement the above processes and system components encompass a broad range from general purpose, programmable computing devices to specialized circuitry, and devices including a combination of both.
  • the processes and system components may be implemented as instructions for computing devices, including general purpose processor instructions for a variety of programmable processors, including microprocessors, Digital Signal Processors, etc. These instructions may be implemented as software, firmware, etc. These instructions can also be converted to various forms of processor circuitry, including programmable logic devices, application specific circuits, including digital, analog and mixed analog/digital circuitry. Execution of the instructions can be distributed among processors and/or made parallel across processors within a device or across a network of devices. Transformation of content signal data may also be distributed among different processor and memory devices.
  • the computing devices used for signal detection and embedding may include, e.g., one or more processors, one or more memories (including computer readable media), input devices, output devices, and communication among these components (in some cases referred to as a bus).
  • processors e.g., one or more processors, one or more memories (including computer readable media), input devices, output devices, and communication among these components (in some cases referred to as a bus).
  • memories including computer readable media
  • input devices e.g., input devices, output devices, and communication among these components (in some cases referred to as a bus).
  • a bus for software/firmware, instructions are read from computer readable media, such as optical, electronic or magnetic storage media via a communication bus, interface circuit or network and executed on one or more processors.
  • the above processing of content signals may include transforming of these signals in various physical forms.
  • Images and video forms of electromagnetic waves traveling through physical space and depicting physical objects
  • the content signals can be transformed during processing to compute signatures, including various data structure representations of the signatures as explained above.
  • the data structure signals in memory can be transformed for manipulation during searching, sorting, reading, writing and retrieval.
  • the signals can be also transformed for capture, transfer, storage, and output via display or audio transducer (e.g., speakers).
  • cell phones While reference has been made to cell phones, it will be recognized that this technology finds utility with all manner of devices - both portable and fixed. PDAs, organizers, portable music players, desktop and laptop computers, tablets, pads, wearable computers, servers, etc., can all make use of the principles detailed herein. Particularly contemplated cell phones include the Apple iPhone, and cell phones following Google's Android specification (e.g., the Gl phone, manufactured for T-Mobile by HTC Corp.). The term "cell phone” should be construed to encompass all such devices, even those that are not strictly- speaking cellular, nor telephones.
  • each includes one or more processors, one or more memories (e.g. RAM), storage (e.g., a disk or flash memory), a user interface (which may include, e.g., a keypad, a TFT LCD or OLED display screen, touch or other gesture sensors, a camera or other optical sensor, a microphone, etc., together with software instructions for providing a graphical user interface), a battery, and an interface for communicating with other devices (which may be wireless, such as GSM, CDMA, W-CDMA, CDMA2000, TDMA, EV-DO, HSDPA, WiFi, WiMax, or Bluetooth, and/or wired, such as through an Ethernet local area network, a T-l internet connection, etc).
  • An exemplary cell phone that can be used to practice part or all of the detailed arrangements is shown in Fig. 1, discussed above.
  • the processor can be a special purpose electronic hardware device, or may be implemented by a programmable electronic device executing software instructions read from a memory or storage, or by combinations thereof.
  • the ARM series of CPUs using a 32-bit RISC architecture developed by Arm, Limited, is used in many cell phones.
  • References to "processor” should thus be understood to refer to functionality, rather than any particular form of implementation.
  • the processor can also comprise a field programmable gate array, such as the Xilinx Virtex series device.
  • the processor may include one or more electronic digital signal processing cores, such as Texas Instruments TMS320 series devices.
  • devices for practicing the detailed methods include operating system software that provides interfaces to hardware devices and general purpose functions, and also include application software that can be selectively invoked to perform particular tasks desired by a user.
  • operating system software that provides interfaces to hardware devices and general purpose functions
  • application software that can be selectively invoked to perform particular tasks desired by a user.
  • Known browser software, communications software, and media processing software can be adapted for uses detailed herein.
  • Some embodiments may be implemented as embedded systems - a special purpose computer system in which the operating system software and the application software is indistinguishable to the user (e.g., as is commonly the case in basic cell phones).
  • the functionality detailed in this specification can be implemented in operating system software, application software and/or as embedded system software.
  • a cell phone communicates with a server at a remote service provider
  • different tasks can be performed exclusively by one device or the other, or execution can be distributed between the devices.
  • description of an operation as being performed by a particular device e.g., a cell phone
  • performance of the operation by another device e.g., a remote server
  • a service provider may refer some tasks, functions or operations, to servers dedicated to such tasks.
  • data can be stored anywhere: local device, remote device, in the cloud, distributed, etc.
  • Operations need not be performed exclusively by specifically- identifiable hardware. Rather, some operations can be referred out to other services (e.g., cloud computing), which attend to their execution by still further, generally anonymous, systems.
  • cloud computing e.g., cloud computing
  • Such distributed systems can be large scale (e.g., involving computing resources around the globe), or local (e.g., as when a portable device identifies nearby devices through Bluetooth communication, and involves one or more of the nearby devices in an operation.)
  • a cell phone may distribute some or all of the image data and/or lighting information to the cloud for analysis, e.g., to detect an encoded signal or to determine image statistics.
  • a detection result, a partial result or computation stages may be communicated back to the cell phone for review or further computation or actions.
  • a detector may only consider one of these considerations when determining color channel weightings.

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Multimedia (AREA)
  • Editing Of Facsimile Originals (AREA)
  • Image Processing (AREA)

Abstract

La détection d'ombres dans des images capturées par des téléphones cellulaires et des caméras en relation avec les tatouages numériques, la dissimulation de données stéganographiques et les empreintes numériques consiste à identifier une ombre projetée par une caméra sur un sujet dont l'image est formée. La détection consiste également à : redresser, au moyen d'un processeur, l'ombre en relation avec la lecture d'un tatouage numérique à partir d'une imagerie capturée du sujet ou le calcul d'une empreinte numérique à partir de l'imagerie capturée du sujet; identifier une ombre projetée par un téléphone cellulaire sur un sujet et, au moyen d'un processeur, déterminer la proximité de la caméra par rapport au sujet sur la base d'une analyse de l'ombre; obtenir des informations d'éclairage associées à une capture d'image ou vidéo et adapter un processus d'identification de signal pour désaccentuer une contribution de signal du canal bleu lorsque les informations d'éclairage sont associées à un éclairage incandescent; et analyser, au moyen d'un processeur, des données d'image ou vidéo pour déduire des informations d'identification de celles-ci. Bien entendu, d'autres revendications et combinaisons sont également proposées.
PCT/US2011/041469 2010-06-23 2011-06-22 Identification et redressement d'ombres en relation avec les tatouages et empreintes numériques; et détection de signal dynamique adaptée sur la base d'informations d'éclairage WO2011163378A1 (fr)

Applications Claiming Priority (4)

Application Number Priority Date Filing Date Title
US35786410P 2010-06-23 2010-06-23
US35791810P 2010-06-23 2010-06-23
US61/357,918 2010-06-23
US61/357,864 2010-06-23

Publications (1)

Publication Number Publication Date
WO2011163378A1 true WO2011163378A1 (fr) 2011-12-29

Family

ID=45371804

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/US2011/041469 WO2011163378A1 (fr) 2010-06-23 2011-06-22 Identification et redressement d'ombres en relation avec les tatouages et empreintes numériques; et détection de signal dynamique adaptée sur la base d'informations d'éclairage

Country Status (1)

Country Link
WO (1) WO2011163378A1 (fr)

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103530867A (zh) * 2013-05-07 2014-01-22 天津大学 一种数码照片闪光灯阴影检测方法
US8983118B2 (en) 2010-06-23 2015-03-17 Digimarc Corporation Determining proximity of a mobile device to a subject based on shadow analysis
EP3099049A1 (fr) * 2015-05-28 2016-11-30 Fujitsu Limited Détection d'information cachée en lumière après réflexion dudite lumière
US10521931B2 (en) 2014-12-19 2019-12-31 Hewlett-Packard Development Company, L.P. Determine image capture position information based on a quasi-periodic pattern
CN115861021A (zh) * 2023-02-23 2023-03-28 深圳市英唐数码科技有限公司 电子书水印修改方法、系统和可读存储介质

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20030112237A1 (en) * 2001-12-13 2003-06-19 Marco Corbetta Method, computer program product and system for rendering soft shadows in a frame representing a 3D-scene
US20040258274A1 (en) * 2002-10-31 2004-12-23 Brundage Trent J. Camera, camera accessories for reading digital watermarks, digital watermarking method and systems, and embedding digital watermarks with metallic inks
US20070104389A1 (en) * 2005-11-09 2007-05-10 Aepx Animation, Inc. Detection and manipulation of shadows in an image or series of images
US20070274561A1 (en) * 1999-05-19 2007-11-29 Rhoads Geoffrey B Methods and devices employing optical sensors and/or steganography
US20090175555A1 (en) * 2008-01-03 2009-07-09 Apple Inc. Illumination systems and methods for computer imagers
US20100119153A1 (en) * 2008-11-13 2010-05-13 Barinder Singh Rai Shadow Remover

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20070274561A1 (en) * 1999-05-19 2007-11-29 Rhoads Geoffrey B Methods and devices employing optical sensors and/or steganography
US20030112237A1 (en) * 2001-12-13 2003-06-19 Marco Corbetta Method, computer program product and system for rendering soft shadows in a frame representing a 3D-scene
US20040258274A1 (en) * 2002-10-31 2004-12-23 Brundage Trent J. Camera, camera accessories for reading digital watermarks, digital watermarking method and systems, and embedding digital watermarks with metallic inks
US20070104389A1 (en) * 2005-11-09 2007-05-10 Aepx Animation, Inc. Detection and manipulation of shadows in an image or series of images
US20090175555A1 (en) * 2008-01-03 2009-07-09 Apple Inc. Illumination systems and methods for computer imagers
US20100119153A1 (en) * 2008-11-13 2010-05-13 Barinder Singh Rai Shadow Remover

Cited By (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8983118B2 (en) 2010-06-23 2015-03-17 Digimarc Corporation Determining proximity of a mobile device to a subject based on shadow analysis
US9336599B2 (en) 2010-06-23 2016-05-10 Digimarc Corporation Determining proximity of a mobile device to a subject based on shadow analysis
CN103530867A (zh) * 2013-05-07 2014-01-22 天津大学 一种数码照片闪光灯阴影检测方法
US10521931B2 (en) 2014-12-19 2019-12-31 Hewlett-Packard Development Company, L.P. Determine image capture position information based on a quasi-periodic pattern
US11263784B2 (en) 2014-12-19 2022-03-01 Hewlett-Packard Development Company, L.P. Determine image capture position information based on a quasi-periodic pattern
EP3099049A1 (fr) * 2015-05-28 2016-11-30 Fujitsu Limited Détection d'information cachée en lumière après réflexion dudite lumière
US9865028B2 (en) 2015-05-28 2018-01-09 Fujitsu Limited Information detecting device, information detecting system, and information detecting method
CN115861021A (zh) * 2023-02-23 2023-03-28 深圳市英唐数码科技有限公司 电子书水印修改方法、系统和可读存储介质

Similar Documents

Publication Publication Date Title
US9336599B2 (en) Determining proximity of a mobile device to a subject based on shadow analysis
US10497081B2 (en) Detecting encoded signals under adverse lighting conditions using adaptive signal detection
US20220335561A1 (en) Embedding signals in a raster image processor
Piva An overview on image forensics
US9489706B2 (en) Device and algorithm for capturing high dynamic range (HDR) video
US11455702B2 (en) Weights embedded to minimize visibility change
US10506128B1 (en) Encoded signal systems and methods to ensure minimal robustness
EP3520390B1 (fr) Recolorisation de flux d'images infrarouges
US9697578B2 (en) Assessment of camera phone distortion for digital watermarking
Gou et al. Intrinsic sensor noise features for forensic analysis on scanners and scanned images
US9070044B2 (en) Image adjustment
CN108241645B (zh) 图像处理方法及装置
US11336795B2 (en) Encoded signal systems and methods to ensure minimal robustness
US10951843B2 (en) Adjusting confidence values for correcting pixel defects
WO2011163378A1 (fr) Identification et redressement d'ombres en relation avec les tatouages et empreintes numériques; et détection de signal dynamique adaptée sur la base d'informations d'éclairage
JP2010211498A (ja) 画像処理プログラム及び画像処理システム
Reed et al. Closed form non-iterative watermark embedding
KR102102403B1 (ko) 위조 영상 인쇄물 감식을 통한 코드인증방법과 그 응용 시스템
CN112541853A (zh) 数据处理方法、装置及设备
US20200074601A1 (en) Hue preservation post processing for highlight recovery
Dua et al. Underwater image enhancement by using amalgamation of colour correction, contrast-enhancing and dehazing

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 11798854

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 11798854

Country of ref document: EP

Kind code of ref document: A1