Due to the reimbursed improvement of web innovation, picture reports have turned into a significant data source. It is difficult to recover certain pictures from every single accessible one. An intuitive picture proposal framework, which initially uses shading histogram highlight and GCLM surface component to express picture substance, at that point a bit based K-implies is used to group pictures into various classes by their visual highlights, at last dependent on an element vectors put away in the database the comparative pictures are recovered. The HSV shading histogram is determined and the joint histogram is inferred dependent on the mix of the tone and immersion in the tint and immersion histogram. The shading highlight is removed from the joint histogram. The chi-square is utilized to discover the likeness between the two pictures. Along these lines worldwide component is determined utilizing the joint histogram. The provincial component is extricated utilizing the GCLM method where the neighbor pixels is considered into record. The assessment results exhibit the exactness of the recovery dependent on the accuracy and review false positive and negative proportion. The ROC bend is utilized to think about the productivity of the shading, surface and the mix of both the shading and the surface.