Abstract: A method and system for detection of contaminants present on a lens of an imaging device is disclosed. An input image received from an imaging device is split into a plurality of patches of predefined size and a kurtosis value calculated for each and compared with a median kurtosis value. Patches having kurtosis value less than the median kurtosis value are selected. Based on comparison of a first maximum likelihood of the selected patches with a predefined threshold, one or more selected patches are stored. Such patches are split into a top and a bottom portion for processing based on discrete wavelet transform and singular value decomposition, respectively. The top and the bottom portion are merged patches for which a second maximum likelihood is greater than a second predefined threshold, are stored. Further, contaminants in the image are classified into predefined categories based on one or more image features.

Abstract: A method and system for detection of contaminants present on a lens of an imaging device is disclosed. An input image received from an imaging device is split into a plurality of patches of predefined size and a kurtosis value calculated for each and compared with a median kurtosis value. Patches having kurtosis value less than the median kurtosis value are selected. Based on comparison of a first maximum likelihood of the selected patches with a predefined threshold, one or more selected patches are stored. Such patches are split into a top and a bottom portion for processing based on discrete wavelet transform and singular value decomposition, respectively. The top and the bottom portion are merged patches for which a second maximum likelihood is greater than a second predefined threshold, are stored. Further, contaminants in the image are classified into predefined categories based on one or more image features.

Abstract: Various embodiments of a method, system and computer program product for correcting distortions in a multimedia content comprising a set of temporal images are disclosed. Each image of the set of temporal images is decomposed into patches. Thereafter, each set of temporal patches is parsed hierarchically to obtain a set of background components and one or more moving objects. Thereafter, the set of background components and the one or more moving objects are aggregated to reconstruct the multimedia content.

Abstract: Various embodiments of a method, system and computer program product for correcting distortions in a multimedia content comprising a set of temporal images are disclosed. Each image of the set of temporal images is decomposed into patches. Thereafter, each set of temporal patches is parsed hierarchically to obtain a set of background components and one or more moving objects. Thereafter, the set of background components and the one or more moving objects are aggregated to reconstruct the multimedia content.