The eye is a complex sensing system. It converts the light stimulus that reaches the retina to electrical signals and also functions as a processor of the data. Understanding the biological visual system has contributed already to major tasks in computerized image processing and machine vision, including medical imaging. In this work, the retinal response to light stimuli at the ganglion cells was investigated and analyzed. A retina of a rat (rattus norvegicus aged 4-6 weeks, after dark adaptation, n=24) was used, and experiments of extracellular recordings in vitro were carried out. We have used isoflurane and CO2 for anesthesia and euthanasia. The eye was enucleated and the retina was extracted and placed on a microelectrode array (MEDA), allowing recording from up to 60 electrodes in parallel. Throughout the recording, the retina was superfused with Ringer’s solution similar to extracellular liquid at 300C. Two sets of experiments were performed. At the first set of experiments, the retina was stimulated with uniform flash illumination. In each stimulus at this set, the pulse duration was 2, 1, 0.5, 0.1 or 0.05 sec. The additional controlled parameter was the dark duration between pulses, as 3.5 or 10 sec. The effect of those parameters on retinal response was analyzed according to three parameters: the response time (between the beginning of the stimulus and the start of the response), the frequency peak (the highest firing rate) and the time-constant of the adaptation process. The second set of experiments was performed with a checker-board pattern stimulus of various resolutions, from a 2X2 checker-board up to a 10X10 checker-board. At both experiments the electrical signals were processed by a dedicated algorithm to improve the signal to noise ratio. For the first set experiments, using uniform illumination, we found that the optimal pulse of 0.1 sec led to maximal firing rate and minimal delay. With longer dark duration, the retina was slower. The time-constant of the adaptation was not affected by the pulse duration and the dark duration. At the second set of experiments, recognition of the checker-board stimulus was successfully performed, such that each recorded point could be characterized as an ON or OFF cell. The resolutions that could be recognized were 2X2 and 3X3. Higher resolutions were not recognized correctly and tended to elicit results closer to those obtained with uniform light. In this experiment, it was found that the number of ON cells was much higher than OFF cells. The conclusions of this research may contribute to the development of artificial bionic vision and in supporting future research in computerized machine vision.