JP2909111B2 - Camera exposure calculation device - Google Patents

Description

Description: BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an exposure calculating device for a camera. 2. Description of the Related Art
As described in Japanese Patent Publication No. 6, a plurality of luminance patterns are set based on the actual image data, each luminance value measured by dividing the screen into a plurality of regions and a luminance pattern are collated, and the center weight is calculated.
Some exposure control is performed by selecting an optimum exposure determining factor from among exposure determining factors such as an average, high luminance emphasis, and low luminance emphasis. Further, as disclosed in Japanese Patent Application Laid-Open No. 61-173226, there is a method in which a backlight state is detected from a spot luminance value at the center of a screen and an average luminance value of the entire screen to perform exposure correction.

[Problems to be solved by the invention]

However, the conventional exposure control can cope with some limited scenes programmed in advance, but cannot cope well with other scenes. In the exposure control that emphasizes the center of the screen, an appropriate exposure can be obtained when there is a subject to be photographed in the center of the screen, but otherwise, an appropriate exposure cannot be obtained. Even in such a case, if the subject you want to shoot is positioned in the center of the screen and then AE-locked, you can always get the proper exposure regardless of the position of the subject, but AE lock is cumbersome to operate. Not suitable for quick shooting.

SUMMARY OF THE INVENTION The present invention has been made in view of the above circumstances, and has as its object to provide an exposure calculation device for a camera having a simple configuration capable of obtaining an optimum exposure in a short time under any circumstances. It is.

[Means and actions for solving the problem]

A focus detection device for a camera according to the present invention includes a single-layer neural network including a plurality of neuron units to which a luminance data vector is input via a predetermined synaptic connection strength. This neural network is trained to output exposure correction value information for a given screen when a brightness data vector, which is a set of brightness data for each of a plurality of points on an arbitrary screen, is input. To determine the exposure. Here, the range of the exposure correction value is about ± 5 EV. At the time of learning, the exposure correction value is assumed to be the same between screens having similar luminance data vectors, and the neural network learns each luminance data vector using a number of model patterns. .

FIG. 1 shows an outline of a means for determining an exposure correction value using a neural network according to the present invention. The input screen shown on the left is a photographing screen shown in the viewfinder, and is divided into regions Pi (i = 1 to n (n is a positive integer, 9 in FIG. 1)) arranged in a two-dimensional matrix, and each region Pi The luminance data Di is obtained every time. The neural network shown on the right side inputs a luminance data vector Di for each region of the input screen and determines an exposure correction value Ex for the screen. The neural network used in the present invention is a single-layer network composed of a plurality of (here, 25) neuron units to which a luminance data vector is input via a predetermined synaptic connection strength. The full luminance data vector is input to the neuron unit.

Learning a neural network consists of the following two steps. First, luminance data vectors in a number of model patterns are input to a neural network,
A predetermined synapse connection strength is obtained by unsupervised learning. Unsupervised learning is the modification of the synaptic connection strength of each neuron unit so that it selectively responds to a given input vector without supervision data. Automatically categorized above. Next, the output neuron unit of the neural network when the luminance data vector of the model pattern is input again to the neural network is associated with the exposure correction value in the model pattern. That is, the value described for each neuron unit in FIG. 1 is the exposure correction value.

At the time of photographing, as shown in FIG. 1, when a luminance data vector Di for each region Pi in the screen to be photographed is input to the neural network trained in this way, one of the neuron units becomes the most important. Ignite strongly (1st
In the figure, the firing neuron unit is shaded),
It can be determined that the exposure correction value Ex (E7 in FIG. 1) corresponding to the neuron unit is a proper exposure correction value, and if the exposure is corrected accordingly, a photograph can always be taken with the proper exposure.

The neural network used in the present invention will be described with reference to FIGS. The neural network model was proposed by Kohonen of Helsinki University of Technology (Finland),
Called a self-organizing network. The learning method is known as self-organizing feature mapping.

FIG. 2 shows a model of each neuron unit. The neuron unit is connected to each input value of the input vector via a synapse connection, compares the input value with each synapse connection strength, and generates an output according to the degree of matching between the two.

The output yi of the ith neuron unit is given by

Where f () is the output function of the neuron unit (usually a sigmoidal monotonically increasing function), wij is the synaptic connection strength for the jth input xkj of the ith neuron unit, and xkj is the kth J is the j-th input value of the input vector, and n is the number of dimensions of the input vector.

The neural network is configured by arranging each neuron unit two-dimensionally. Here, FIG. 3 shows a one-dimensionally modeled one for simplicity of explanation.

As shown in FIG. 3, the neural network performs a self-organizing feature mapping that introduces the effect of feedback coupling of the signal from the output to the input of the neuron unit. Here, wik is the feedback coupling strength from the k-th neuron unit to the i-th neuron unit.

FIG. 4 shows the degree of the influence of such feedback coupling on the synaptic coupling strength. In other words, the synaptic connection strength shows a change in the characteristics of the Mexican hat type shape due to the effect of the feedback connection, and when the output value of a specific neuron unit increases, the neuron units in a positional relationship close to the distance also draw to it. Therefore, the output value of the neuron unit outside the output value decreases.

Simplifying the learning of the synapse connection strength taking into account the effect of this feedback connection so that it is performed only near the optimally matched neuron unit, the learning process (self-organizing feature mapping) of this neural network is as follows. The procedure can be performed as follows.

# 1: Wi (10) is initialized with random numbers, and the number of times of learning t = 0. Here, Wi (t) = (wi1 (t), wi2 (t),... W
in (t)).

Hereinafter, the processing of steps # 2 and # 3 is repeated for each input vector Xk.

# 2: Assuming that t = t + 1, an optimal combined neuron unit c satisfying the following equation is obtained.

{Xk−wc (t)} = min {Xk−Wi (t)} (2) where Xk = (xki, xk2,..., Xkn), and wc (t) is the optimal combined neuron unit c. It is the synaptic connection strength.

# 3: If iｔNe (t), Wi (t + 1) = Wi (t) + α
(T) × (Xk−Wi (t)), and if i∈Nc (t), Wi (t + 1) = Wi (t).

Here, α (t) is a learning coefficient, and Nc (t) is a set of neuron units near the optimal combined neuron unit c. Normally, both α (t) and Nc (t) are monotonically decreasing functions. It is.

By performing the self-organizing feature mapping in this manner, the synaptic connection strengths of the neighboring neuron units become similar. In addition, since synapse connection strength reflecting statistical properties such as similarity and correlation between model patterns is obtained, vector quantization of each model pattern is performed, and classification of the model patterns becomes possible. Further, the process of determining the exposure correction value of an arbitrary photographing screen can be considered as the process of determining the optimal combined neuron unit c. Is possible.

〔Example〕

Hereinafter, an embodiment of an exposure calculation device for a camera according to the present invention will be described with reference to FIGS. 5 to 7. FIG. 5 is a block diagram of an automatic exposure control camera as an embodiment. A plurality of light-receiving elements 2i (i = 1 to n) for detecting the luminance of each of a plurality of regions of the subject on one screen incident through the lens 1 and the light-receiving elements 2i (i = 1 to n) are provided.
Is supplied to the normalization operation circuit 3. The normalization operation circuit 3 normalizes the luminance data Di to a real value from 0 to 1 so that it can be input to the neural network 5, and obtains the input data pattern xkj described above. The output of the normalization operation circuit 3 is input to a single-layer neural network 5 having a synaptic connection strength already acquired by learning and comprising a plurality of neuron units.

The neural network 5 receives the luminance data xkj, and receives the synaptic connection strength w of each neuron unit.
distance between ij (t) and vector , And the minimum value of the distance between the vectors Dc = min ｛Ni｝
And outputs an exposure correction value Ex corresponding to the neuron unit giving the minimum value. In general, the neural network 5 has a total number of photometric areas of 5 to 5.
It consists of about 20 times as many neuron units. This exposure correction value information is input to the exposure calculation circuit 4. Open F-number, aperture value, shutter speed, film sensitivity, mode information (shutter priority, aperture priority, etc.) of the lens from the shooting information input unit 8
Is also supplied to the exposure calculation circuit 4. The exposure calculation circuit 4 performs an apex calculation (brightness value−open F value + film sensitivity = shutter speed + aperture value) and an exposure correction calculation according to the mode information, and outputs the result to the shutter control circuit 6 and the aperture control circuit 7. Supply.

FIG. 6 shows a specific circuit configuration of the neural network 5. The neural network 5 includes memories 9, 10, 11, and 12 for storing various data and various arithmetic circuits 13 and 14.
Consists of The memory 9 stores the normalized luminance data xkj output from the normalization operation circuit 3, the memory 10 stores the synapse coupling strength wij (t) obtained by learning, and the synapse coupling with the normalized luminance data xkj. A memory 11 for storing a distance Ni between vectors with the intensity wij (t),
Neurons on the trained neural network
A memory 12 is provided for storing exposure correction value information indicating the correspondence between the unit and the exposure correction value of the input pattern. As an arithmetic circuit, a circuit that calculates a distance Ni between vectors from the normalized luminance data xkj and the synaptic connection strength wij (t).
13 and a circuit 14 for detecting the minimum value of the vector distance Ni
Is provided.

Normalized luminance data x output from the normalization operation circuit 3
kj is stored in the memory 9 in the neural network 5. Synapse connection strength wij obtained by learning
(T) is stored in the memory 10 in advance. Memory 9, 10
Is supplied to the inter-vector distance calculation circuit 13, where the inter-vector distance is calculated. Can be performed, and the calculation result is stored in the memory 11. Since the calculation of the distance between vectors is independent for each neuron unit, parallel calculation is possible. Therefore, the inter-vector distance calculation circuit 13 is composed of a plurality of processors that perform processing for each neuron unit, and can perform high-speed processing. Further, in the arithmetic element dedicated to the neural network, that is, in the neurochip, each arithmetic processing is performed by each neuron unit.

The minimum value detection circuit 14 detects an optimal combined neuron unit c having the minimum value from the obtained vector distance Ni. The exposure calculation circuit 4 reads out the exposure correction value information Ex corresponding to the optimally-suited neuron unit from the correspondence relationship predetermined and stored in the memory 12, and obtains the exposure-correction value information corresponding to the obtained optimally-suited neuron unit. Apex calculation and exposure correction calculation are performed based on Ex and the open F-number, aperture value, shutter speed, film sensitivity, and mode information supplied from the shooting information input unit 8, and the results are used as the shutter control circuit 6 and the aperture control device 7. To supply.

As described above, automatic exposure control appropriate for the scene of the subject in the input screen can be performed.

As the exposure correction value information, the expression (1) defining the output of the neuron unit is: It is also possible to use an exposure correction value obtained by weighting and averaging a plurality of exposure correction values corresponding to a plurality of neuron units selected in order from the maximum value of the values obtained in accordance with the output value of the neural network. In addition, the neural network in the case of applying only the learned result is possible only with the above-described configuration. However, in order to perform the initial learning and the additional learning when adapting the neural network to the camera operator, the following is required. The learning circuit configuration described in (1) is required.

FIG. 7 illustrates a circuit configuration for learning a neural network. This configuration is similar to the circuit shown in FIG. 6 except that the synapse connection strength calculation circuit 16 and the exposure correction value information setting circuit 18
Is realized simply by adding

In the operation, first, the synapse connection strength wij of each neuron unit of the neural network is initialized with a small random number. The luminance data of each region for a large number of learning model patterns is stored in the memory 15, and the data Ex relating to the exposure correction value is stored in the memory 17.
After determining the optimum combined neuron unit by detecting the minimum value of the inter-vector distance using the luminance data of the memory 15 as in the case of FIG.
Learns the synaptic connection strength Wi of the neighboring neuron unit group centered on the optimal combined neuron unit by the following equation. Here, i∈Nc (t).

Wi (t + 1) = Wi (t) + α (t) × (Xk−Wi (t)) (3) where α (t) is a learning coefficient and takes a real value of 0 to 1, and Nc (t) Is a set of neuron units near the optimal combined neuron unit c, the initial value of which is not so small as compared with the size of the neural network, and both α (t) and Nc (t) are monotonically decreasing functions. It is. t is the number of times of learning, and the learning is repeated until the maximum number of times of learning tmax.

After completing the learning for all model patterns,
The exposure correction value information setting circuit 18 checks which exposure correction value of each model pattern stored in the memory 17 is associated with which neuron unit of the neural network according to the following relationship, and stores the result in the memory 12. Set as exposure compensation value information.

In the case of (threshold value in the k-th input pattern), the exposure correction value of the k-th input pattern is made to correspond to the neuron unit i.

In the case of performing additional learning to adapt to the camera operator in the trained neural network, when an arbitrary input pattern is given, the optimal matching neuron unit is determined by detecting the minimum value of the distance between vectors. , A neuron that is in the vicinity of the neuron unit and has a common exposure correction value.
The synapse connection strength Wi of the unit group Nc 'is corrected and learned by the synapse connection strength calculation circuit 16 according to the following equation.

 Wi = Wi + αo × (Xk−Wi) (4) where i∈Nc ′ and αo is a learning coefficient.

The present invention is not limited to the above-described embodiment, but can be variously modified. The learning method described is unsupervised learning. When an input pattern is given, the output of the neural network is compared with the teacher data (exposure correction value of the input pattern). If the neural network fires correctly, Bring the synapse connection strength closer to the input pattern,
In addition, supervised learning that keeps the synapse connection strength away from the input pattern if the fire is erroneously performed is also possible. Also, learning can be performed including teacher data as input values. In order to be able to return to the initial learning state before additional learning at any time, it is necessary to store the synapse connection strength in the initial learning state in a memory.

Further, as input data to be self-organized by the neural network, not only the luminance data of the entire screen described above, but also the luminance data of the in-focus position and the vicinity thereof, and the luminance data of the main subject and the vicinity thereof in cooperation with the automatic focusing mechanism. Brightness data may be used.

Although the embodiment has described the automatic exposure control camera, the present invention may be applied to a manual exposure control camera that only displays the determined exposure information in a viewfinder or the like.

〔The invention's effect〕

As described above, according to the present invention, in a process that has conventionally been difficult to formulate and program, such as determination of an exposure correction value according to the scene of a subject, a neural network has a large number of model patterns. By simply learning the luminance data vector, it is possible to provide an exposure calculation device for a camera that can automatically and quickly obtain exposure correction values for various screens. If the parallel processing of the neural network is used, the speed can be further increased. Also, by letting the neural network learn the camera operator's favorite composition and composition habit and adapt the neural network to the camera operator, the exposure according to the photographer's intention can be automatically determined.

[Brief description of the drawings]

FIG. 1 is a diagram schematically showing an exposure calculation device for a camera according to the present invention, FIG. 2 is a diagram showing a model of a neuron unit, FIG. 3 is a diagram showing a model of a neural network, and FIG. FIG. 5 is a block diagram of an embodiment of an automatic exposure control camera as an embodiment of the present invention, FIG. 6 is a diagram showing a circuit configuration of a neural network, and FIG. FIG. 3 is a diagram showing a circuit configuration for learning a neural network. DESCRIPTION OF SYMBOLS 1 ... Lens 2 ... Light receiving element 3 ... Normalization calculation circuit 4 ... Exposure calculation circuit 5 ... Neural network 6 ... Shutter control circuit 7 ... Aperture control circuit
8 ... photographing information input unit.

Claims (2)

(57) [Claims] 1. A luminance detecting means for obtaining a luminance data vector which is a set of luminance data corresponding to the luminance of each of a plurality of regions of a subject image incident through a photographing lens at the time of photographing; A large number of neuron units connected to each of the data vectors via the respective synaptic connection strengths are formed in a single-layer network, and when a luminance data vector corresponding to the model pattern is input in advance, Each synaptic connection strength is learned so that the unit responds selectively,
A neural network for providing a value related to exposure correction based on a state of the neuron unit selectively reacting to the luminance data vector at the time of the imaging by the single-layer network having the plurality of neuron units; Exposure control means for controlling the exposure of the camera in accordance with the value relating to the exposure correction generated by (1). 2. Acquiring the synaptic connection intensity by modifying the synaptic connection intensity so as to selectively react to the luminance data vector of the model pattern, and re-transmitting the luminance data vector of the model pattern to the neural network. 2. The exposure calculation device for a camera according to claim 1, further comprising learning means for associating an output of the neural network when input with the exposure correction value of the model pattern.