Changelog for package object_recognition_tod

0.5.6 (2016-04-24)

• fix OpenCV3 compilation
• Contributors: Vincent Rabaud

0.5.5 (2015-04-19)

• Fix seg fault while detecting objects In many cases, matches[i] contains less than 5 elements (varies seemingly random between 0 and 4) Long history available [here](https://github.com/plasmodic/ecto_opencv/commit/2f2fe7fb75d09337c1d594cee416bd948f337b30#commitcomment-10687068)
• Contributors: Jorge Santos Simón

0.5.4 (2015-04-15)

• fix radius search not support with LSH
• remove useless code
• Contributors: Vincent Rabaud

0.5.3 (2015-03-31)

• use OpenCV and not opencv_candidate for LSH
• Contributors: Vincent Rabaud

0.5.2 (2015-03-30)

• add opencv_candidate as a dependency It is included but just transitively so it's better to have it directly
• Revert "remove dependency on opencv_candidate" This reverts commit 3a515e48c815b9e0c7f6df20b730b98d0dccd5df. The RGBD module from opencv_candidate is indeed needed
• Contributors: Vincent Rabaud

0.5.1 (2015-03-29)

• remove dependency on opencv_candidate
• clean extensions
• remove useless build dependency
• Contributors: Vincent Rabaud

0.5.0 (2014-04-13)

• compile under Indigo
• Merge pull request #6 from awesomebytes/patch-2 For some reason this is needed using openni
• For some reason this is needed using openni This is just a workaround, the real problem is that the default values expect: `[ INFO] [1395248076.002277778]: Subscribed to topic:/camera/rgb/camera_info with queue size of 1 [ INFO] [1395248076.003073369]: Subscribed to topic:/camera/rgb/image_raw with queue size of 1 [ INFO] [1395248076.003968179]: Subscribed to topic:/camera/depth_registered/camera_info with queue size of 1 [ INFO] [1395248076.004994459]: Subscribed to topic:/camera/depth_registered/image_raw with queue size of 1` Using openni2 this is not a problem as it's published by default but using openni those are not published, so this workarounds that.
• Merge pull request #5 from awesomebytes/patch-1 Configuration that gives detections
• Configuration that gives detections Tested with an Asus Xtion and 6 objects trained. The sink2 + adding to pipeline1 the output sink2 makes that /recognized_object_array has publications!
• fix tests
• add openni runtime dependency
• fix the ROS config file to use ROS
• drop Fuerte support
• update url
• udpate the docs
• Merge pull request #3 from destogl/ros_cell_config_correction Corrected cell cofing to work out-of-the-box with ROS.
• Corrected cell cofing to work out of the box with ros. There was problem with wrong type and parameters cell is added.
• fixes wg-perception/object_recognition_core#13
• Contributors: Denis Štogl, Sammy Pfeiffer, Vincent Rabaud

0.4.16 (2013-05-29)

• use proper K_image/K_depth
• update email address
• Contributors: Vincent Rabaud

0.4.15 (2013-04-14)

• fix display
• fix the 3d point non-sense
• comply to the new core API
• finally convert the Trainer fully to C++
• remove old split modules in favor of one cell
• comply to the new DB API
• implement moreof the pipeline
• comply to the new headers in core
• add a Trainer cell to replace the Python mess
• make the cell contents be in a common library-like file
• fix the pipeline
• fix the invalid plasm
• Contributors: Vincent Rabaud

0.4.14 (2013-03-12)

• comply to the new DB API

File truncated at 100 lines see the full file