Submit Search
Representation learning by learning to count
3 likes
1,159 views
Fujimoto Keisuke
Representation learning by learning to count
Technology
Read more
1 of 18
Download now
Downloaded 23 times
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
More Related Content
PDF
A quantum computational approach to correspondence problems on point sets
Fujimoto Keisuke
PDF
F0-Consistent Many-to-many Non-parallel Voice Conversion via Conditional Auto...
Fujimoto Keisuke
PDF
YOLACT real-time instance segmentation
Fujimoto Keisuke
PDF
Product Managerの役割、周辺ロールとの差異
Fujimoto Keisuke
PDF
ChainerRLで株売買を結構頑張ってみた(後編)
Fujimoto Keisuke
PDF
Temporal Cycle Consistency Learning
Fujimoto Keisuke
PDF
ML@Loft
Fujimoto Keisuke
PDF
20190414 Point Cloud Reconstruction Survey
Fujimoto Keisuke
A quantum computational approach to correspondence problems on point sets
Fujimoto Keisuke
F0-Consistent Many-to-many Non-parallel Voice Conversion via Conditional Auto...
Fujimoto Keisuke
YOLACT real-time instance segmentation
Fujimoto Keisuke
Product Managerの役割、周辺ロールとの差異
Fujimoto Keisuke
ChainerRLで株売買を結構頑張ってみた(後編)
Fujimoto Keisuke
Temporal Cycle Consistency Learning
Fujimoto Keisuke
ML@Loft
Fujimoto Keisuke
20190414 Point Cloud Reconstruction Survey
Fujimoto Keisuke
More from Fujimoto Keisuke
(20)
PDF
Chainer meetup 9
Fujimoto Keisuke
PDF
20180925 CV勉強会 SfM解説
Fujimoto Keisuke
PDF
Sliced Wasserstein Distance for Learning Gaussian Mixture Models
Fujimoto Keisuke
PDF
LiDAR-SLAM チュートリアル資料
Fujimoto Keisuke
PDF
Stock trading using ChainerRL
Fujimoto Keisuke
PDF
Cold-Start Reinforcement Learning with Softmax Policy Gradient
Fujimoto Keisuke
PDF
Dynamic Routing Between Capsules
Fujimoto Keisuke
PDF
Deep Learning Framework Comparison on CPU
Fujimoto Keisuke
PDF
ICCV2017一人読み会
Fujimoto Keisuke
PDF
Global optimality in neural network training
Fujimoto Keisuke
PDF
CVPR2017 oral survey
Fujimoto Keisuke
PDF
Point net
Fujimoto Keisuke
PDF
CVIM最先端ガイド6 幾何学的推定のための最適化手法 3.5 - 3.8
Fujimoto Keisuke
PDF
Value iteration networks
Fujimoto Keisuke
PDF
sublabel accurate convex relaxation of vectorial multilabel energies
Fujimoto Keisuke
PDF
Deep SimNets
Fujimoto Keisuke
PDF
G2o
Fujimoto Keisuke
PDF
Deep Fried Convnets
Fujimoto Keisuke
PDF
L0TV: a new method for image restoration in the presence of impulse noise
Fujimoto Keisuke
PDF
Learning to forget continual prediction with lstm
Fujimoto Keisuke
Chainer meetup 9
Fujimoto Keisuke
20180925 CV勉強会 SfM解説
Fujimoto Keisuke
Sliced Wasserstein Distance for Learning Gaussian Mixture Models
Fujimoto Keisuke
LiDAR-SLAM チュートリアル資料
Fujimoto Keisuke
Stock trading using ChainerRL
Fujimoto Keisuke
Cold-Start Reinforcement Learning with Softmax Policy Gradient
Fujimoto Keisuke
Dynamic Routing Between Capsules
Fujimoto Keisuke
Deep Learning Framework Comparison on CPU
Fujimoto Keisuke
ICCV2017一人読み会
Fujimoto Keisuke
Global optimality in neural network training
Fujimoto Keisuke
CVPR2017 oral survey
Fujimoto Keisuke
Point net
Fujimoto Keisuke
CVIM最先端ガイド6 幾何学的推定のための最適化手法 3.5 - 3.8
Fujimoto Keisuke
Value iteration networks
Fujimoto Keisuke
sublabel accurate convex relaxation of vectorial multilabel energies
Fujimoto Keisuke
Deep SimNets
Fujimoto Keisuke
G2o
Fujimoto Keisuke
Deep Fried Convnets
Fujimoto Keisuke
L0TV: a new method for image restoration in the presence of impulse noise
Fujimoto Keisuke
Learning to forget continual prediction with lstm
Fujimoto Keisuke
Ad
Representation learning by learning to count
4.
• - • - - • • -
5.
• - • - • - • •
6.
• - - - • - →気持ちを画像に掛けて、それを直接学習 →こういう気持ちが画像に掛かってるだろう、 という仮説に基づいて学習
7.
•
8.
•
9.
• - - - - -
11.
• 0.34 0.54 0.02 0.41 1.31 CNN CNN CNN CNN
12.
• • 1.31 1.21 ここを⼀ 致させる
13.
• - • - - • • •
14.
•
15.
• - - • • • -
16.
• - - Highest counting Lowest counting
17.
• - •
18.
• • • •
Download