Garment Modeling From A Single Image

Modeling garment was very useful in various fields.Most recent garment remodeling techniques rely on acquiring multiple views of clothing,which may not always available,especially in the case of photographs from the web.For these reasons,we propose a method that is able to compute a rich reality 3D model of clothing from a single photograph with little human interaction.Some effective algorithms for garment modeling with a single image was proposed and implemented in this paper.Without extra input,quality garment models would be generated by this platform.A Pedestrian detection based on SVM and HOG was used to gain the human area of the Image.The garment silhouetted was extracted by Grabcut method from the human area.The garment silhouetted was processed by a garment parser based on SVM((Support Vector Machine)and KNN(K-NearestNeighbor).By extending the Asynchronous Contact Mechanics algorithm,our algorithm got the garment patterns according to the garment parsing result.The garment patterns were matched with reality texture,our algorithm triangulated the extracted contours by Bowyer-Watson triangulation.According to the nearest point go together principle,as-rigid-as-possible was used to reshape the garment contours.At last our garment patterns had reality texture.In addition our algorithm parametric changed the garment patterns.our algorithm also proposed Computer-typesetting algorithms for garment patterns.