Merge pull request #6 from Roboy/bohare-patch-1

Update 01_solution_strategy.rst

docs/development/01_solution_strategy.rst

@@ -3,16 +3,17 @@ Solution Strategy
 
 Basic decisions for Vision Package:
 
-- Seperation of different tasks into sub-modules (Face Detection, Object tracking, face recognition, mood recognition, age estimation, scene classification, ...)
+- Seperation of different tasks into sub-modules (Face Detection, Object Detection, Object tracking, face recognition, mood recognition, age estimation, scene classification, ...)
 - Highest priority on face detection and face pose estimation. Recognition of people as second priority. All other properties concerning people with lower priority and general object detection with least importance.
 - Face detection using this approach: `Joint Face Detection and Alignment using MTCNNs <https://kpzhang93.github.io/MTCNN_face_detection_alignment/paper/spl.pdf>`_. Good real time performance, other modules to be built on top.
 - Face embeddings using `FaceNet <https://arxiv.org/pdf/1503.03832.pdf>`_. These embeddings can be used for recognition.
-
+- Object Detection using 'You Only Look Once 9000 <https://arxiv.org/abs/1612.08242>'.  
 
 Current implementation:
 
 - **Face detection** running in realtime with function to query whether a face is nearby (to be called as ROS service).
 - **Face recognition** implemented as extra thread (to be called as ROS service on a given face). Using Facenet for calculating embeddings for a given face and SVM for classification. SVM currently trained on pictures of LFW (labelled Faces in the Wild) dataset, using Roboy Team members as next step.
+- **Object Detection** is currently working on images; capable of detecting objects with 76.8 mAP on VOC 2007 dataset.
 - **ROS services** currently not implemented in devel branch. Only workaround in dirty_final_hack branch.
 
 Plan for coming semester with priorities in red (5 being highest priority):