Techfolio

Trending New Technologies Demos by Camelia Ciolac

Camelia Follow Camelia

MediaPipe FaceMesh TFLite for Mascarons 3D Reconstruction

In this mini-projet, we use Google’s MediaPipe framework as well as a pretrained Tensorflow Lite model to predict the face mesh given an image of a mascaron (sculptured keystone). We visualize the reconstructed face in 3D, with skin texture.

27 Nov 2020

« TF Instance Segmentation and Dominant Colors of Kimonos OpenCV Salient Points and Local Descriptors to Recognize Landmark Building in Century-old Postcard »

comments powered by Disqus

Explore →

Tensorflow (7) CENTER-NET (1) Human Pose Keypoints (1) OpenCV (3) Instance Tracking (1) FFMPEG (1) Computer Vision (8) MASK R-CNN (1) Instance Segmentation (1) HDBSCAN (1) ColorThief (1) TFLite (1) FaceMesh (1) 3D (1) Pyvista (1) image salient points (1) image local descriptors (1) RANSAC (1) similarity (3) Siamese Net (1) ResNet50 (3) Embeddings (2) Lambda Layer (4) Custom Callback (2) TF Dataset (2) Triplet Net (1) Custom Loss (1) Annoy index (1) DeepLab (1) Image Segmentation (1) Confidence heatmap (1) webscraping (1) MongoDB (1) declarative ML (1) custom Callbacks (1) Keras Tuner (1) tf.data (1) tf.feature_column (1) tabular data (1) Optunity (1) CatBoost (1) Image classification (1) NLP (2) Spacy (2) AllenNLP (1) WordNet (1) BabelNet (1) BERT (1) Node2Vec (1) SupWSD (1) Part-of-speech (1) NetworkX (1) SPARQL (1) RDF (1) Blazegraph (1) GCP (2) NER (1) Word2Vec (1) Gensim (1) Doccano (1)