Image processing device, an image processing method, and computer-readable recording medium
An image processing device according to one of the exemplary aspects of the present invention includes: a scale space generation means for generating the scaled samples from a given input region of interest; feature extraction means for extracting features from the scale samples; a likelihood estimation means for deriving an estimated probability distribution of the scaled samples by maximizing the likelihood of a given scaled sample and the parameters of the distribution; a probability distribution learning means for updating the model parameters given the correct distribution of the scaled samples; a template generation means to combine the previous estimates of the object features into a single template which represents the object appearance; an outlier rejection means to remove samples which have a probability below the threshold; and a feature matching means for obtaining the similarity between a given template and a scaled sample and selecting the sample with the maximum similarity as the final output.
1. An image processing device comprising:
a feature extraction unit that extracts features from scaled samples generated from given region of interest, after normalizing the samples;
a maximum likelihood estimation unit that derives an estimated probability score of the scaled samples by maximizing the likelihood of a given scaled sample and a parameter of the probability distribution model;
an estimation unit that combines the previous estimates of the object and its features into a single template which represents the object appearance, and that removes samples which have a probability score below the threshold;
a feature matching unit that obtains a similarity between a given template and a scaled sample and selecting the sample with the maximum similarity as the final output.
2. The image processing device according to claim 1 , further comprising
a learning unit that updates the probability distribution model parameters given the distribution of the scaled samples and the template derived from the previous frames.
3. The image processing device according to claim 1 ,
Wherein the maximum likelihood estimation unit obtains the probability that a sample is generated by distribution which is given by the model of the distribution of the features, the model is applied to the newly generated scale samples and a score is calculated based on the distance of the samples.
4. The image processing device according to claim 2 ,
Wherein the learning unit that learns the probability distribution models parameters by one or more series of training samples and template which are given as true samples and generated from the previous frames.
5. The image processing device according to claim 1 ,
Wherein estimation unit that combines the previous estimates of the object and its features into a single template which represents the object appearance.
6. An image processing method comprising:
a step (a) of extracting features from scaled samples generated from given region of interest, after normalizing the samples;
a step (b) of deriving an estimated probability distribution score of the scaled samples by maximizing the likelihood of a given scaled sample and a parameters of the probability distribution model;
a step (c) of combining the previous estimates of the object and its features into a single template which represents the object appearance;
a step (d) of removing samples which have a probability score below the threshold;
a step (e) of obtaining a similarity between a given template and a scaled sample and selecting the sample with the maximum similarity as the final output.
7. The image processing method according to claim 6 , further comprising
a step (f) of updating the probability distribution model parameters given the distribution of the scaled samples and the template derived from the previous frames.
8. The image processing method according to claim 6 ,
Wherein in the step (b), obtaining the probability that a sample is generated by distribution which is given by the model of the distribution of the features, the model is applied to the newly generated scale samples and a score is calculated based on the distance of the samples.
9. The image processing method according to claim 7 ,
Wherein in the step (f) learning the probability distribution models parameters by one or more series of training samples and template which are given as true samples and generated from the previous frames.
10. The image processing method according to claim 6 ,
Wherein in the step (c) combining the previous estimates of the object and its features into a single template which represents the object appearance.
11. A non-transitory computer-readable recording medium storing a program that causes a computer to operate as:
a feature extraction unit that extracts features from scaled samples generated from given region of interest, after normalizing the samples;
a maximum likelihood estimation unit that derives an estimated probability score of the scaled samples by maximizing the likelihood of a given scaled sample and a parameters of the probability distribution model;
an estimation unit that combines the previous estimates of the object and its features into a single template which represents the object appearance, and that removes samples which have a probability score below the threshold;
a feature matching unit that obtains a similarity between a given template and a scaled sample and selecting the sample with the maximum similarity as the final output.
12. The non-transitory computer-readable recording medium according to claim 11 , further the program causes the computer to operate as:
a learning unit that updates the probability distribution model parameters given the distribution of the scaled samples and the template derived from the previous frames.
13. The non-transitory computer-readable recording medium according to claim 11 ,
Wherein the maximum likelihood estimation unit obtains the probability that a sample is generated by distribution which is given by the model of the distribution of the features, the model is applied to the newly generated scale samples and a score is calculated based on the distance of the samples.
14. The non-transitory computer-readable recording medium according to claim 12 ,
Wherein the learning unit that learns the probability distribution models parameters by one or more series of training samples and template which are given as true samples and generated from the previous frames.
15. The non-transitory computer-readable recording medium according to claim 11 ,
Wherein estimation unit that combines the previous estimates of the object and its features into a single template which represents the object appearance.