The online clothing industry via e-commerce websites is looking forward to expanding its profit by producing realistic and well-fitted outfits as per the body appearance. Preserving these clothing details and designs, i.e., logo, embroidery, makes it more realistic. Image-based virtual try-on network requires a model that can accept as input a 2D consumer image and a garment, and then produce an image of the consumer dressed in the garment. In the past, some research have been done to preserve this clothing detailing, but characteristics preservation suffered from distortion and misalignment. Though we have better generators now, we need to encourage for the preservation of the characteristics. In this paper, we use a deep learning framework with two residual-based encoders along with a decoder followed by fusion blocks to achieve warped clothing and target segmentation. The fusion block contains residual blocks, in which we arrange the convolution layers and dropouts in such a way that it preserves the clothing details. The experimental finding demonstrates that the suggested model yields better outcomes than conventional approaches in terms of SSIM score.