Fish is such a useful food with a great nutritional value that has the most of essential nutrients for human. In recent decades, due to population growth and under-nourishment of half of the world’s people, fish can be considered as an important source of protein supplement. In foodstuffs, the quality of a product includes all the features that indicate its nutritional value. Fish freshness is an important index for its quality. According to the nutritional value of fish in the food basket, achieving new technologies for its freshness assessment is of great importance. In this research, 20 farmed rainbow trout ( Oncorhynchus mykiss ) fishes were prepared and their eyes and gills color changes were monitored by using machine vision technique in order to evaluate their freshness. A digital color imaging system was employed to record the visual characteristics of the eyes and gills of the samples. To provide the desired and similar optical conditions, a light chamber was designed and fabricated, and different optical conditions were tested to acquire the best images. By collecting the photos of the samples for 10 days, 80 colored photos were acquired from the right and left gills and eyes during the time period of 1 to 2 pm. The camera settings, as well as the location of the camera lens were the same on all days of imaging. The R, G, and B were determined as redness, greenness, and blueness parameters during ice-storage of the samples. The region of interest was automatically selected using a computer program developed in MATLAB software. The initial recalled images in MATLAB software were in the RGB color space. Then, the color photos were transferred to the computer and processed in image processing toolbox of the software. Before the main processing operation, a pre-processing step was conducted on the images in order to remove the noise. Afterwards, the R, G, and B components were extracted from the initial images and the final black and white images were obtained through applying the threshold and eliminating undesired small points. Finally, by applying the logical AND operation, the final colored images were obtained. The analysis of variance results indicated that keeping days of the samples had a significant effect on all color components (RGB) of both eye and gill, and by passing the time their color was getting whiter. In next step, 54 color features were extracted in the RGB, HSV, and L*A*B* color spaces. Artificial neural network (ANN) and support vector machines (SVMs) were utilized for Keywords: ltr"
