1. QT-NET (git)
   • Finished the report of experiment 6 about illumination.
   • Make an experiment plan about the number of images in each illumination, and do a part of the experiments.
2. SLAM
   • Two methods are used train the network: (1) all pixels, (2) consistent pixels.
   • The pose estimation can be improved with the combination of (1) and (2). … Continue Reading ››
     

1. QT-NET
   • Finished the experiments about the illumination, include 9, 5, 3, 1 kinds of illumination. The result is upload to git.
2. SLAM
   • continue to do some experiments about the consistency of pixel.

1. QT-NET
   • Finished the 1.0 version of the research report, and upload to the git. But, some grammar error and express error need to be modified in the future.
   • The experiment about 3 group of illumination is finished today, next week I will try to test the network with 1 group of illumination. Due … Continue Reading ››
     

• About the positive and negative group, I do some experiments to test the influence of illumination. For 9, 5, 3 group of illumination, I compared the accuracy and loss. And all the experiment is upload to git.
• The result is that maybe illumination just has a litter influence to the accuracy.
• Continue Reading ››
  

1. QT-Net
   • About the positive and negative group, I do some experiments to test the influence of times of each object are used for combination.
   • 
2. SLAM
   • I am doing some work to add the occlusion mask, consistence mask, and consistence score … Continue Reading ››
     

参加：M2全員

いぎみ：楽天市場のAPI連動テストを行う

水戸：人工知能関連アルゴリズムに関する研究をテーマとしてのテーマの決定。

財前：進行中の学習済みモデルで制度を検証してみる

北原：被験者６名のデータを集めて、制度を評価。

二石：４種類の標識を用いて、検証用画像とラベリングを行い、制度を評価する。

• About the result of past experiments, I make a summary. I think it may be because the data set is too simple, causing the network to be over-fitted before it is well-fitted. The reasons why I think so are :
  1. only the positive face is used for train;
  2. The result that training with … Continue Reading ››
     

• Do some experiments to test the influence of some parameters, including the number of illumination type, the number of image in each illumination, and the difference between illuminations.
• About the distribution of each image in combination, making a combination about the mask of object instead of rectangle include object. This modification make the … Continue Reading ››