Added video, validation set and ground truth for tracking module

Added David testdata (Images, ground truth and metadata)

- Removed old David dataset
- Added README

Added Dudek testdata (Images, ground truth and metadata)

Corrected first rect of David ground truth

Added the initOmit files with frames where the target is occluded or out
of view

Added faceocc2 testdata (Images, ground truth and metadata)

Added perf xml

Removed sequences of images and added video files for david, dudek and faceocc2 dataset

Author: lenlen

testdata/cv/tracking/README.md

@@ -0,0 +1,16 @@
+## Testdata for tracking module
+
+All image files and ground truth are taken from.
+
+The original images sequences are combined into video files with this command:
+ffmpeg  -i %4d.jpg -vcodec libvpx -crf 4 -b 1M testset.webm
+
+```
+@inproceedings{ WuLimYang13,
+  Title = {Online Object Tracking: A Benchmark},
+  Author = {Yi Wu and Jongwoo Lim and Ming-Hsuan Yang},
+  Booktitle = {IEEE Conference on Computer Vision and Pattern Recognition (CVPR)},
+  Year = {2013}
+}
+```
+WEB SITE: https://sites.google.com/site/trackerbenchmark/benchmarks

testdata/cv/tracking/david/data/david.webm

@@ -0,0 +1,5 @@
+%YAML:1.0
+att: [ 1, 1, 1, 1, 1, 1, 0, 1, 0, 0, 0 ]
+start: 300
+prefix: ""
+suffix: "jpg"