Assignment 2

Assignment 2

project-2018/requirement_update_1.md

@@ -0,0 +1,60 @@
+# Real World Data Analysis
+
+We have deployed your code on the robot. And the robot has diven the following real data.
+
+| sensor 1 input | sensor 2 input |
+| ===============|================|
+|[(knife,1, 99%), (scissor, 2, 65%), (spoon, 3, 33%), (spoon, 4, 80%), (keys, 5, 95%)] | [ (keys, 5, 95%), (spoon, 4, 99%),(fork, 3, 99%), (scissor, 2, 95%),  (knife,1, 55%)]|
+| [empty list] | [empty list] |
+| [(knife,1, 99%), (scissor, 2, 65%), (spoon, 3, 33%)] | [empty list] |
+| [(knife,1, 99%), (scissor, 2, 65%), (spoon, 3, 33%)] | [(KNIFE,1, 99%), (SCISSOR, 2, 65%), (SPOON, 3, 33%)] |
+| [(knife,1, 99%), (scissor, 2, 65%)] | [(fork, 3, 99%), (spoon, 4, 99%)] |
+| [(knife,1, 94%),(knife,1, 69%),(knife,1, 89%)]|[(knife,1, 99%),(fork, 3, 99%)] |
+
+## Task
+* Please test you code with the following input and report the output.
+* If the code didnot run for the following output report the case for the failure, update the code and run the test.
+
+
+# Increase of input
+
+This week we found out that we can add 2 new deep learning algortihms has been found to be showing state-of-the art performance
+for object recognition tasks.
+1. Image matching algorithm [1]
+2. RGB-D Object Detection and Semantic Segmentation for Autonomous Manipulation in Clutter [2]
+
+So now we have info from 4 different input sources:
+| Sensor 1 (algo 1) | Sensor 1 (algo 2)|  sensor 2 (algo 1) | sensor 2 (algo 2)|
+|===================|==================|====================|==================|
+|[(knife, 1, 89%)]  | [(knife, 1, 35%)] | [(knife, 1, 69%)] | [(knife, 1, 80%)]|
+
+## Task
+
+* Update your code for the following new information.
+* Does your code follow the "S.O.L.I.D" principle of "OPen for update but close for changes" ? If yes how ?
+* How to test the "Open for update" principle ? You should only add new lines of code and update least line of code.
+* Another test for principle? If next time a modality needs to be removed then by commenting out some lines of code should suffice.
+* If it doesnt follow the principle , find which design pattern will help? Refactor your code for the design pattern.
+* Write test cases.
+* Create an csv sheet or table to mention all the test cases you are trying .
+
+## Points
+
+    * 10 Working in group
+    * 10 Working with issues in github (converting requirements to github issues, commiting to them and closing them on completion)
+    * 10 Working with branch and using pull request to update the master code(No commit to master code)
+    * 30 TDD appraoch
+    * 20 Using Design Patterns
+    * 20 Solution
+
+
+
+
+
+## References
+[1] Robotic  Pick-and-Place  of  Novel  Objects  in  Clutter
+with  Multi-Affordance  Grasping  and  Cross-Domain  Image  Matching
+[2] Max Schwarz, Anton Milan, Arul Selvam Periyasamy, and Sven Behnke:
+RGB-D Object Detection and Semantic Segmentation for Autonomous Manipulation in Clutter
+Accepted for International Journal of Robotics Research (IJRR), Sage Publications, to apprear 2017.
+Available at: http://ais.uni-bonn.de/papers/IJRR_2017_Schwarz.pdf