INDIAN-CELEBNET: A Dataset for Generative Modelling of Indian Celeb Facesdoi:10.1109/i2ct45611.2019.9033631Animesh KarnewarChinmay SawantIEEEInternational Conference for Convergence for Technology
The CelebFace+ dataset contains 202, 599 face images of 10, 177 celebrities. People in CelebFaces+ and LFW are claimed to be mutually exclusive. The CASIA- WebFace [16] is currently the largest dataset which is publicly available, with about 10K celebrities, and 500K images. A quick ...
we also provide a very large training dataset to facilitate the task. The training dataset contains about 10M images for 100K top celebrities selected from our one-million celebrity list in terms of their web appearance frequency. Our training data is, to the best of our knowledge, the largest...
We develop a community detection based pipeline to clean the noisy MS-Celeb-1M face dataset. As the diversity of faces is preserved in multiple large communities, our cleaning results have both high cleanness and rich data diversity. More details can be found in our paper here. The picture be...
However, video samples produced by CogVideo [26] shown in Figure 6 are of a lower quality than the ones by MMVID [19] trained solely on CelebV-Text, where gener- ated faces are not in a high relevance to input texts, demon- strating the effectiveness of our facial text-v...
Some images are very blurry and even clearly not human faces. We provide a wash list to clean the dataset, and you can download from OneDrive(coming soon) or Baidu Yun. Some Details This list is based on XiangWu's repo, so it will may be slightly cleaner. The ID mapping is same as...
We develop a community detection based pipeline to clean the noisy MS-Celeb-1M face dataset. As the diversity of faces is preserved in multiple large communities, our cleaning results have both high cleanness and rich data diversity. More details can be found in our paper here. The picture be...