# MotionCNN

**Repository Path**: cedaroski/MotionCNN

## Basic Information

- **Project Name**: MotionCNN
- **Description**: No description available
- **Primary Language**: Python
- **License**: Not specified
- **Default Branch**: master
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 0
- **Forks**: 0
- **Created**: 2026-04-08
- **Last Updated**: 2026-04-08

## Categories & Tags

**Categories**: Uncategorized

**Tags**: None

## README

# MotionCNN
## NOT WORK YET
It is a supervised learning method for camera ego-motion estimation.
A sub-mission for a SLAM system.
## Dependency
tensorflow 1.10

## train
`
python moton_train.py --mode train --data_path /home/user/Data/2019-03-12/ --filenames_file /home/user/Data/2019-03-12/binocular_train.txt --log_directory tmp --model_name train2layer --num_epochs 100 --batch_size 1
`
## TODO
1. read the ego-motionl label
2. Superivised learning of ego-motion
3. By the depth prediction from monodepth, change the net to unsuperivised