Читать книгу Recognition and Perception of Images - Группа авторов - Страница 28

The work of [Buhmann et al., 1999] is dedicated to the methods of recognition of image and scenes in case of computer vision. Our visual system can reconstruct the three-dimensional figures on the basis of a flat image. The computer vision requires not only the recognition algorithms. Additionally, it is necessary to consider the lighting, shadows and scene perspective. Our perception of the surrounding world allows us to exclude the ambiguities. It is offered to apply the gestalt rules of objects grouping to recognize the scenes at the computer vision in the work. The images are to be analyzed from the whole to the part at the initial level while finding out the connection and interaction of constituent parts. Thereafter, the analyzed information is specified and interpreted. Furthermore, the objects are classified and recognized.

The visual field surrounding us consists of objects of various shapes and colors; taking into account the distance, some overlap others, some are perceived as a figure, some as a background. From an evolutionary point of view, it is important for us to recognize the necessary object against the background of the environment [Shiffman, 2008], [Aliyeva, 2007]. Usually, out of two objects of the field of view, a region with a marked border is perceived as a figure; it is smaller in size, or differs in filling, and is located closer to the observer (Figure 1.3.1) [Abbasov, 2019]. The fragments perceived as a figure contain for us more interesting information. Relationship figure and background can be perceived not only visually, but also in other sensory systems. For example, you can hear a bird singing in the background of street noise or a violin melody in the background of the sounds of the rest of the orchestra.

Figure 1.3.1 The combination of “figure-background”.

The process of perception of three-dimensional objects on the basis of photographs and paintings is in fact an illusion. Our visual system is adapted to the perception of the real world based on the signs of the depth and distance of the picture plane.