EP4587962A1 - Verfahren und vorrichtung zum trainieren eines objektdetektors mit einem ensemble von mehreren detektionsköpfen - Google Patents
Verfahren und vorrichtung zum trainieren eines objektdetektors mit einem ensemble von mehreren detektionsköpfenInfo
- Publication number
- EP4587962A1 EP4587962A1 EP23768504.5A EP23768504A EP4587962A1 EP 4587962 A1 EP4587962 A1 EP 4587962A1 EP 23768504 A EP23768504 A EP 23768504A EP 4587962 A1 EP4587962 A1 EP 4587962A1
- Authority
- EP
- European Patent Office
- Prior art keywords
- neural network
- head
- training
- image
- output
- Prior art date
- Legal status (The legal status 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 status listed.)
- Withdrawn
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N5/00—Computing arrangements using knowledge-based models
- G06N5/04—Inference or reasoning models
- G06N5/045—Explanation of inference; Explainable artificial intelligence [XAI]; Interpretable artificial intelligence
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/0464—Convolutional networks [CNN, ConvNet]
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
Definitions
- a neural network which is trained according to a method with features of claim 1, enables the advantages of object detection using an ensemble (more precise predictions, more precise uncertainty determination), while the computing time is only slightly increased.
- the invention relates to a computer-implemented method for training a neural network, wherein the neural network is designed to detect objects in images, and the training comprises the following steps: Determining a feature representation of a training image using a backbone of the neural network; ⁇ Determining head outputs based on the feature representation and by means of a plurality of heads of the neural network, the head outputs each characterizing a presence or absence of an object for areas of the training image or characterizing an extent of the object or characterizing a position of the object or the extent and the position characterized; ⁇ Determining a plurality of first loss values, a first loss value corresponding to a head output and characterizing a deviation of the head output from a desired head output; ⁇ Adjusting at least one parameter of the spine and/or at least one parameter of a head
- So sensor in particular optical sensors can be used, in particular a camera.
- images within the meaning of the invention can also be recorded by other sensors, for example a LIDAR, a radar, an ultrasonic sensor or a thermal camera.
- the output signal can also characterize a spatial extent of at least one object within the image and/or a positioning of the object within the image.
- the output signal can also characterize a spatial expansion with respect to real dimensions of the object in the real world and/or positioning within a coordinate system of the real world.
- the intrinsic and extrinsic parameters of a camera can be used to assign a coordinate in a predefined coordinate system of the real world to the pixels and thus objects of an image.
- the neural network can then be trained in such a way that instead of determining the position and/or extent of the object with respect to a width and height of the image in pixels, a position and/or extent of the object with respect to the predefined coordinate system.
- the neural network itself can be understood as consisting of two parts.
- the first part of the network includes the backbone.
- backbone is common in the state of the art.
- base or base network are occasionally used synonymously with the term backbone.
- the backbone sometimes referred to as a feature extractor, can in turn be understood as a neural network.
- the task of the backbone is to transform the image into the feature representation, on the basis of which object detection can then be carried out.
- the well-known CenterNet architecture there is a head for determining the presence/absence, a head for determining the position and a head for determining the width and height of the object.
- the well-known Single Shot Detection (SSD) architecture there are two heads, one of which carries out the presence/absence determination and the other head carries out the position and extent determination.
- the structure of a Faster RCNN looks similar to the SSD.
- the neural network presented in this invention determines a plurality of heads to determine the absence/absence or to determine the position or to determine the extent or to determine the extent and the position.
- the heads can therefore be understood as an ensemble, which is formed, the presence/absence or the position or the extent or the position R.401853 - 6 - and to determine the extent.
- the inventors were able to find that the introduction of an ensemble leads to a significant increase in the accuracy of object detection.
- a head is preferably a neural network.
- the neural network can therefore preferably be composed of several sub-networks, namely the backbone and the heads.
- other models from the field of machine learning are also possible as backbones and heads, for example support vector machines.
- the heads of the plurality of heads each determine a head output. For training, these head outputs are compared with a desired head output in order to determine a first loss value.
- a plurality of heads can also be provided for determining the presence/absence and for determining the position and for determining the extent.
- This can be understood as meaning that there is an ensemble for determining the presence/absence, the position and the extent in the neural network, i.e. a plurality of pluralities of heads.
- the inventors were able to determine that the accuracy of the neural network can be increased even further by using a plurality of heads.
- the invention relates to executing the trained neural network, i.e.
- the head outputs determined by the majority of heads are combined into the one head output, with the output signal which characterizes an object detection of objects in the image being determined based on this combined head output.
- Various methods can be chosen for summarizing.
- the head outputs can be the same in their form (for example matrix or tensor) (for example the same R.401853 - 9 - Number of dimensions and elements in the respective dimensions). This can be the case in particular if the individual heads are the same in their architecture (for example, the same number of layers with the same number of neurons) or the last layers of the heads are the same.
- the average values of elements at the same index positions of the header outputs can be used for summarization.
- each header output can be assigned a weight, which is used for all values of the corresponding header output when forming the weighted sum.
- the inventors were able to determine that the special design of the network through a backbone and the ensemble of heads advantageously increases the accuracy of the network. In addition, the requirement for computing capacity and memory is significantly lower compared to using a deep ensemble.
- the output signal additionally includes a value that characterizes an uncertainty regarding the summary of the plurality of head outputs.
- the user is thus advantageously given the opportunity to intervene in expenditures that are associated with a high degree of uncertainty in order to reduce or suppress possible misconduct.
- the head outputs are typically in the form of a matrix or a tensor
- the uncertainty can be determined in particular with respect to each element or a subset of elements of the matrix or tensor.
- This plurality of uncertainties can then be provided in the output signal.
- the user of the neural network can be shown which area of the image the object detection has determined which uncertainty.
- the uncertainty can be expressed as a value that is determined using a predefined function. In particular, mutual information can be used as a function.
- an entropy can be determined for each header output and a maximum entropy can be provided as an uncertainty value with respect to the index position for each index position within the header outputs.
- a method is provided for determining a control signal of an actuator or a display device based on an image, wherein the control signal is determined based on the output signal.
- the actuator can be controlled better in this way, since the output signal enables more precise object detection.
- the control signal can in particular also be determined based on the determined uncertainty.
- FIG. 1 shows a neural network (60) which is set up for training.
- the neural network (60) includes a backbone ( ⁇ ), which in turn characterizes a neural network, for example a vision transformer.
- the backbone is designed to determine a feature representation ( ⁇ ) from a training image ( ⁇ ⁇ ), which the neural network receives as input.
- This feature representation is sent to a plurality ( ⁇ ⁇ ) of heads ( ⁇ ⁇ , ⁇ ⁇ , ⁇ ⁇ ) of the neural network (60) as input of the heads ( ⁇ ⁇ , ⁇ ⁇ , ⁇ ⁇ ) transmitted.
- the heads ( ⁇ ⁇ , ⁇ ⁇ , ⁇ ⁇ ) are preferably also neural networks, for example convolutional neural networks (CNN).
- CNN convolutional neural networks
- the heads are set up to determine the presence or absence of objects.
- the heads ( ⁇ ⁇ , ⁇ ⁇ , ⁇ ⁇ ) each determine head outputs ( ⁇ ⁇ , ⁇ ⁇ , ⁇ ⁇ ) in the form of matrices whose elements correspond to areas of the training image ( ⁇ ⁇ ).
- the receptive field of an element of the matrix defines the extent of the area within the training image ( ⁇ ⁇ ). An element can then characterize whether there is an object within the area of the training image corresponding to it or not.
- an element can indicate whether a center of a bounding box that encloses the object lies within the region.
- the element can preferably be a real-valued number between 0 and 1, which can be characterized as the probability of the existence of an object. From a predefined threshold value, for example 0.5, an element can then be understood as indicating the presence of an object. In this case, values less than 0.5 characterize the absence of objects within the corresponding area.
- the threshold value can be understood as a hyperparameter of the neural network.
- the neural network (60) comprises a further plurality ( ⁇ ⁇ ) of heads ( ⁇ ⁇ , ⁇ ⁇ , ⁇ ⁇ ) to determine an extent of the object (for example width and height within the image).
- the heads ( ⁇ ⁇ , ⁇ ⁇ , ⁇ ⁇ ) of this further plurality ( ⁇ ⁇ ) determine further head outputs ( ⁇ ⁇ , ⁇ ⁇ , ⁇ ⁇ ), each of which characterizes the exact position and in the output signal ( ⁇ ⁇ ) to be provided.
- the extent and position are determined jointly by heads of a plurality of heads.
- the desired output signal ( ⁇ ⁇ ) preferably includes a desired head output for the presence or absence of objects within the training image ( ⁇ ⁇ ), for a position of objects within the training image ( ⁇ ⁇ ) and for an extent of the objects within of the training image ( ⁇ ⁇ ).
- the neural network (60) determines an output signal ( ⁇ ⁇ ) based on the input image ( ⁇ ⁇ ).
- the desired output signal ( ⁇ ⁇ ) and the determined output signal ( ⁇ ⁇ ) are transmitted to a change unit (180). Based on the desired output signal ( ⁇ ⁇ ) and the determined output signal ( ⁇ ⁇ ), the change unit (180) then determines new parameters ( ⁇ ′) for the neural network (60).
- the change unit (180) first determines a plurality of first loss values, each loss value being a deviation of a head output ( ⁇ ⁇ , ⁇ ⁇ , ⁇ ⁇ , ⁇ ⁇ , ⁇ ⁇ , ⁇ ⁇ , ⁇ ⁇ , ⁇ ⁇ , ⁇ ⁇ , ⁇ ⁇ , ⁇ ⁇ , ⁇ ⁇ , ⁇ ⁇ , ⁇ ⁇ ⁇ ) of the output signal ( ⁇ ⁇ ) is characterized by a corresponding desired head output included in the desired output signal ( ⁇ ⁇ ).
- the head outputs ( ⁇ ⁇ , ⁇ ⁇ , ⁇ ⁇ , ⁇ ⁇ , ⁇ ⁇ , ⁇ ⁇ , ⁇ ⁇ , ⁇ ⁇ , ⁇ ⁇ , ⁇ ⁇ , ⁇ ⁇ , ⁇ ⁇ , ⁇ ⁇ ) included in the output signal ( ⁇ ⁇ ) can each be in the form of a matrix or a three-dimensional tensor are present, in which case the desired head output encompassed by the desired output signal ( ⁇ ) is also present in the form of a matrix or a three-dimensional tensor.
- the corresponding matrices or tensors each contain values along a height and a width dimension R.401853 - 15 - or vectors that characterize a presence or absence, a position or an extent.
- Values or vectors at the same indices along the height and width dimensions of a head output ( ⁇ ⁇ , ⁇ ⁇ , ⁇ ⁇ , ⁇ ⁇ , ⁇ ⁇ , ⁇ ⁇ , ⁇ ⁇ , ⁇ ⁇ , ⁇ ⁇ , ⁇ ⁇ , ⁇ ⁇ , ⁇ ⁇ ⁇ ) and a corresponding head output can be are then compared using a loss function.
- the values determined in this way can be accumulated using a mathematical operation, for example summed, weighted summed or their maximum value can be determined.
- ⁇ ( ⁇ ) , ... , ⁇ ( ⁇ ) is the plurality of head outputs
- ⁇ is the number of head outputs
- ⁇ is the number of elements along the height dimension within a head output
- ⁇ is the number of elements along the width dimension within a head edition
- ⁇ is the Euclidean norm.
- the determined new parameters ( ⁇ ′) are provided to the neural network (60) as parameters ( ⁇ ).
- the training described is repeated iteratively for a predefined number of iteration steps or repeated iteratively until the first loss value falls below a predefined threshold value.
- the training is ended when an average first loss value with respect to a test or validation data set falls below a predefined threshold value.
- the new parameters ( ⁇ ′) determined in a previous iteration are used as parameters ( ⁇ ) of the neural network (60).
- the neural network (60) comprises a summary unit ( ⁇ ⁇ , ⁇ ⁇ , ⁇ ⁇ ) at the inference time, which Header issues of the majority of header issues are combined into a single header issue ( ⁇ , ⁇ , ⁇ ).
- the presence or absence, position and extent are determined based on a respective plurality of head outputs.
- one summary unit ( ⁇ , ⁇ , ⁇ ) is provided for each plurality of head outputs in the neural network (60).
- FIG. 4 shows a control system (40), which is designed to determine a control signal (A) of an actuator (10) based on the neural network (60).
- the control system (40) receives the sequence of sensor signals (S) from the sensor (30) in an optional receiving unit (50), which converts the sequence of sensor signals (S) into a sequence of input images ( ⁇ ) converted (alternatively, the sensor signal (S) can also be taken over directly as an input image ( ⁇ ).
- the input image ( ⁇ ) can be, for example, a section or further processing of the sensor signal (S). In other words, the input image ( ⁇ ) is determined depending on the sensor signal (S).
- the sequence of input images ( ⁇ ) is fed to the neural network (60).
- the neural network (60) is preferably parameterized by parameters ( ⁇ ) which are stored in a parameter memory (DB) and are provided by it.
- control system (40) comprises at least one processor (45) and at least one machine-readable storage medium (46) on which instructions are stored which, when executed on the at least one processor (45), Cause the control system (40) to carry out the method according to the invention.
- a display device (10a) is provided as an alternative or in addition to the actuator (10).
- the display device (10a) can in particular be an optical display device, for example a screen.
- the display device (10a) can also be set up for other human senses. For example, it can be a loudspeaker or display an event to a user via haptic signals such as vibration.
- the control signal (A) can then be determined in such a way that the actuator or actuators (10) are controlled in such a way that the motor vehicle (100), for example, prevents a collision with the objects identified by the neural network (60), in particular if These are objects of certain classes, e.g. pedestrians.
- the display device (10a) can be controlled with the control signal (A).
- the display device can be, for example, a screen on which the identified objects are displayed. It is also possible for the display device (10a) to be controlled with the control signal (A) in such a way that it emits a visual or acoustic warning signal if it is determined that the motor vehicle (100) is threatening to strike one of the identified objects collide.
- the output signal ( ⁇ ) includes information that characterizes the identity of a person detected by the neural network (60), and the control signal (A) is selected based on the identity of the person.
- logical access control can also be provided.
- Figure 8 shows an exemplary embodiment in which the control system (40) is used to control a monitoring system (400). This exemplary embodiment differs from the exemplary embodiment shown in FIG. 7 in that instead of the actuator (10), the display device (10a), which is controlled by the control system (40), is provided.
- the sensor (30) is set up to determine an image of a patient, for example an X-ray image, an MRI image or an ultrasound image. At least part of the image is transmitted to the neural network (60) as an input image ( ⁇ ).
- the neural network (60) can, for example, be set up to detect different types of tissue to be recognized on the input image ( ⁇ ).
- the control signal (A) can then be selected such that the determined types of tissue are highlighted in color on the display device (10a).
- the imaging system (500) can also be used for non-medical purposes, for example to determine material properties of a workpiece. For this purpose, the imaging system (500) can record an image of a workpiece.
- the neural network (60) can be set up in such a way that it receives at least part of the image as an input image ( ⁇ ) and detects materials of the workpiece.
- the detection determined in this way can, for example, be displayed together with the input image on the display device (10a).
- the term “computer” includes any device for processing predetermined calculation rules. These calculation rules can be in the form of software, or in the form of hardware, or even in a mixed form of software and hardware.
- a plurality can be understood as indexed, that is, each element of the plurality is assigned a unique index, preferably by assigning consecutive integers to the elements contained in the plurality.
- a plurality comprises ⁇ elements, R.401853 - 24 - where ⁇ is the number of elements in the plural, the elements are assigned the integers from 1 to ⁇ .
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Computing Systems (AREA)
- Software Systems (AREA)
- Artificial Intelligence (AREA)
- Mathematical Physics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Biomedical Technology (AREA)
- Molecular Biology (AREA)
- General Health & Medical Sciences (AREA)
- Biophysics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Image Analysis (AREA)
Abstract
Description
Claims
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| DE102022209517.0A DE102022209517A1 (de) | 2022-09-12 | 2022-09-12 | Verfahren und Vorrichtung zum Trainieren eines Objektdetektors |
| PCT/EP2023/074544 WO2024056503A1 (de) | 2022-09-12 | 2023-09-07 | Verfahren und vorrichtung zum trainieren eines objektdetektors mit einem ensemble von mehreren detektionsköpfen |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| EP4587962A1 true EP4587962A1 (de) | 2025-07-23 |
Family
ID=88018238
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| EP23768504.5A Withdrawn EP4587962A1 (de) | 2022-09-12 | 2023-09-07 | Verfahren und vorrichtung zum trainieren eines objektdetektors mit einem ensemble von mehreren detektionsköpfen |
Country Status (4)
| Country | Link |
|---|---|
| EP (1) | EP4587962A1 (de) |
| CN (1) | CN119923649A (de) |
| DE (1) | DE102022209517A1 (de) |
| WO (1) | WO2024056503A1 (de) |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US11334766B2 (en) * | 2019-11-15 | 2022-05-17 | Salesforce.Com, Inc. | Noise-resistant object detection with noisy annotations |
| US12456047B2 (en) * | 2019-11-21 | 2025-10-28 | Google Llc | Distilling from ensembles to improve reproducibility of neural networks |
| DE102020209162A1 (de) * | 2020-07-21 | 2022-01-27 | Robert Bosch Gesellschaft mit beschränkter Haftung | Verfahren und Vorrichtung zum Detektieren von Anomalien in Sensoraufzeichnungen eines technischen Systems |
-
2022
- 2022-09-12 DE DE102022209517.0A patent/DE102022209517A1/de active Pending
-
2023
- 2023-09-07 WO PCT/EP2023/074544 patent/WO2024056503A1/de not_active Ceased
- 2023-09-07 CN CN202380065242.6A patent/CN119923649A/zh active Pending
- 2023-09-07 EP EP23768504.5A patent/EP4587962A1/de not_active Withdrawn
Also Published As
| Publication number | Publication date |
|---|---|
| WO2024056503A1 (de) | 2024-03-21 |
| CN119923649A (zh) | 2025-05-02 |
| DE102022209517A1 (de) | 2024-03-14 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| DE102020210352A1 (de) | Verfahren und Vorrichtung zum Transferlernen zwischen modifizierten Aufgaben | |
| EP3853778B1 (de) | Verfahren und vorrichtung zum betreiben eines steuerungssystems | |
| DE102020212515A1 (de) | Verfahren und Vorrichtung zum Trainieren eines maschinellen Lernsystems | |
| DE202019105282U1 (de) | Vorrichtung zum Optimieren eines System für das maschinelle Lernen | |
| DE102023213094A1 (de) | Verfahren und Vorrichtung zur Validierung von Erklärbarkeitsmethoden für ein maschinelles Lernsystem | |
| DE102020208828A1 (de) | Verfahren und Vorrichtung zum Erstellen eines maschinellen Lernsystems | |
| DE102018218834A1 (de) | Verfahren und Vorrichtung zum Ermitteln eines Ansteuersignals | |
| DE102022201679A1 (de) | Verfahren und Vorrichtung zum Trainieren eines neuronalen Netzes | |
| DE102020212514A1 (de) | Verfahren zur Ermittlung eines Ausgabesignals mittels eines maschinellen Lernsystems | |
| DE102021209212A1 (de) | Verfahren und Vorrichtung zum Ermitteln von Objektdetektionen eines Bildes | |
| EP4587962A1 (de) | Verfahren und vorrichtung zum trainieren eines objektdetektors mit einem ensemble von mehreren detektionsköpfen | |
| DE102018211875A1 (de) | Verfahren und Vorrichtung zum Betreiben eines Steuerungssystems | |
| DE102021108907A1 (de) | Verfahren und Vorrichtung zum Ermitteln einer Klassifikation und/oder eines Regressionsergebnisses unter fehlenden Sensordaten | |
| DE102020212108A1 (de) | Verfahren und Vorrichtung zum Anlernen eines maschinellen Lernsystems | |
| WO2020173700A1 (de) | Verfahren und vorrichtung zum betreiben eines steuerungssystems | |
| DE102020211262A1 (de) | Verfahren und Vorrichtung zum Komprimieren eines neuronalen Netzes | |
| DE102021202342A1 (de) | Verfahren und Vorrichtung zum Trainieren eines Klassifikators und/oder Regressors | |
| DE102022205824A1 (de) | Verfahren und Vorrichtung zum Trainieren eines neuronalen Netzes | |
| DE202020104727U1 (de) | Vorrichtung zum Transferlernen zwischen modifizierten Aufgaben | |
| DE102018216298A1 (de) | Verfahren und Vorrichtung zum Betreiben eines Steuerungssystems | |
| DE102021209643A1 (de) | Verfahren und Vorrichtung zur Objektdetektion und/oder zur semantischen Segmentierung | |
| DE102021207937A1 (de) | Verfahren und Vorrichtung zum Erstellen eines maschinellen Lernsystems mit einer Mehrzahl von Ausgängen | |
| DE102018216295B4 (de) | Verfahren und Vorrichtung zum Ermitteln eines Ansteuersignals zum Ansteuern eines Aktors | |
| DE102020208309A1 (de) | Verfahren und Vorrichtung zum Erstellen eines maschinellen Lernsystems | |
| DE102020213527A1 (de) | Verfahren zum Optimieren einer Strategie für einen Roboter |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| STAA | Information on the status of an ep patent application or granted ep patent |
Free format text: STATUS: UNKNOWN |
|
| STAA | Information on the status of an ep patent application or granted ep patent |
Free format text: STATUS: THE INTERNATIONAL PUBLICATION HAS BEEN MADE |
|
| PUAI | Public reference made under article 153(3) epc to a published international application that has entered the european phase |
Free format text: ORIGINAL CODE: 0009012 |
|
| STAA | Information on the status of an ep patent application or granted ep patent |
Free format text: STATUS: REQUEST FOR EXAMINATION WAS MADE |
|
| 17P | Request for examination filed |
Effective date: 20250414 |
|
| AK | Designated contracting states |
Kind code of ref document: A1 Designated state(s): AL AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HR HU IE IS IT LI LT LU LV MC ME MK MT NL NO PL PT RO RS SE SI SK SM TR |
|
| DAV | Request for validation of the european patent (deleted) | ||
| DAX | Request for extension of the european patent (deleted) | ||
| STAA | Information on the status of an ep patent application or granted ep patent |
Free format text: STATUS: THE APPLICATION IS DEEMED TO BE WITHDRAWN |
|
| 18D | Application deemed to be withdrawn |
Effective date: 20251028 |