The fashion industry has been having a significant influence on many other fields. In which identification and classification of clothes play a crucial role. It is applied in e-commerce systems to search and recommend suitable outfits for customers, track and capture trending outfits for suitable business solutions, etc. In this study, we present an approach using transfer learning and segmentation techniques to identify types of clothes. We compare and evaluate the classification results of some transfer learning approaches such as InceptionV3, AlexNet, and Transfer Forest. First, we have modified the ImageNet dataset by filtering out mislabeled data. Then we performed data enhancement by changing a few training metrics and increasing the number of training images to 320,000 images, including 45∘∘, 90∘∘, 180∘∘ rotated images of the images from the refined dataset. The experimental results have an accuracy of 0.51 with the dataset that we have modified and can reach 0.97 in accuracy in three-class classification tasks. Finally, we have performed segmentation of clothing images with Unet for supporting further analysis on detailed characteristics of clothes.