发明名称 Image compression apparatus, image expansion apparatus, and methods and programs thereof
摘要 An image compression apparatus capable of compressing an input image that includes a predetermined target object at a high compression ratio while allowing high quality image restoration. In the apparatus, the input image is reduced and compressed. A region of interest corresponding to a predetermined target object is set in the input image. A partial area image of an expanded image of the compressed reduced image corresponding to the region of interest is converted to a high resolution image by applying a prediction process that uses a learning result obtained by learning the predetermined object in advance. With respect to the region of interest portion, a differential image between the image converted to the high resolution and the input image is generated and encoded. Reduced image compression data and differential image encoded data are outputted.
申请公布号 US8805105(B2) 申请公布日期 2014.08.12
申请号 US201213457025 申请日期 2012.04.26
申请人 FUJIFILM Corporation 发明人 Takamori Tetsuya;Kameyama Hirokazu
分类号 G06K9/36 主分类号 G06K9/36
代理机构 Birch, Stewart, Kolasch & Birch, LLP 代理人 Birch, Stewart, Kolasch & Birch, LLP
主权项 1. A non-transitory computer-readable medium storing a program, the program when executed performs image compression using at least one processor which comprises: an image reducing unit that reduces an input image having a first resolution at a predetermined reduction ratio to generate a reduced image having a second resolution lower than the first resolution; a reduced image compression unit that compresses the reduced image to generate reduced image data; a reduced image expansion unit that expands the reduced image data; a region of interest setting unit that sets a region of interest corresponding to a predetermined target object in the input image; a high resolution conversion unit that converts a partial image of an area of the expanded reduced image corresponding to the region of interest to a partial image having the first resolution by applying a prediction process that uses a learning result obtained by learning the predetermined object in advance; a differential image generation unit that generates a differential image by taking a difference between the partial image converted to the first resolution and a partial image of the region of interest portion of the input image; a differential image encoding unit that encodes the differential image to generate differential image data; and an output unit that outputs the reduced image data and the differential image data.
地址 Tokyo JP