Uses of video-based spectroscopy are expanding in many domains where depth information is crucial, including 3D vision, medical diagnostics, distant object identification, and space reconnaissance. For accurate image processing and to enhance the authenticity, depth information is vital in all of these cases. Saliency detection is a useful technique in this method for predicting which parts of images or movies people would find most interesting. However, stereoscopic visual saliency has received surprisingly little attention. This study investigates and evaluates the most recent multimodal fusion based picture identification model from stereoscopic video coupled with saliency approaches in an effort to address this deficiency. The results demonstrate that the adaptive multimodal fusion based spectroscopic enabled image identification model with saliency features is superior when compared to an image detection model from stereoscopic movies that does not use saliency approach.