For Visual impaired people it is difficult to choose different types of clothes during online shopping. We are developing system which helps blind people to recognize color of clothes along with different categories such as material, size, patterns etc. our system convert speech into text format and then display the result of different category with speech so that blind people choose the clothes they want. Identification of color is based on histogram of each image in HSI color space and multilevel clustering is used for identification of items that satisfying many local feature. For speech recognition Deep learning method are used. This approach is helpful for blind people as well as handicapped people