DeepHL is a web-based open system for supporting comparative analysis of animal trajectories.
Please refer to the user guide in the supplementary information of our Nature Communications paper. As for requests related to collaboration of behavior analysis using DeepHL, please e-mail to the corresponding author.
ADAX GATEIO
Occupied GPUs: 0/2, Available memory:500GB
Elapsed time from the last keep alive message from the server is 1924 hours. This means that the server is down for approximately 1924 hours. If this situation continues, please e-mail to maekawa@ist.osaka-u.ac.jp.
GATE.IO-MITGLIEDSCHAFT
Takuya Maekawa, Kazuya Ohara, Yizhe Zhang, Matasaburo Fukutomi, Sakiko Matsumoto, Kentarou Matsumura, Hisashi Shidara, Shuhei J. Yamazaki, Ryusuke Fujisawa, Kaoru Ide, Naohisa Nagaya, Koji Yamazaki, Shinsuke Koike, Takahisa Miyatake, Koutarou D. Kimura, Hiroto Ogawa, Susumu Takahashi, and Ken Yoda: Deep Learning-assisted Comparative Analysis of Animal Trajectories with DeepHL, Nature Communications 11 (5316), Oct. 2020.
We explain how a user prepares a zip file containing the trajectory files to be uploaded to DeepHL. First, the user creates a folder for each of the two classes to store trajectory files belonging to the class, i.e., the user creates two folders, for example, ``male'' and ``female'' folders. The folder name is used as a class name in DeepHL. Then, the user stores a CSV file prepared for each trajectory in its corresponding folder, e.g., a file containing a male trajectory is stored in the male folder. Finally, the user creates a zip file containing these folders. (For example, the user compresses the parent folder of these two folders in a zip file.)
We then explain the format of a CSV file containing a trajectory. The first column of the file corresponds to the timestamp (numeric), the second column corresponds to the X coordinate, and the third column corresponds to the Y coordinate. The first line of the file is the header, and each of the following lines describes the timestamp, X coordinate, and Y coordinate at each time slice. The time units of the timestamp are also used in DeepHL to calculate features such as speed. Please include a header in the CSV file because DeepHL uses a name of a column when additional sensor data is included in CSV.
For more detail, see the supplementary information of our Nature Communications paper.
GATE IO BANKÜBERWEISUNG
Frequent format errors:
"__MACOSX" directory should be removed in a zip file for upload.
Time stamps are duplicated.
Coordinate values are missing in some lines.
Time should be numeric.
Other TIPS:
It is better to use more than 50 trajectories in total because DeepHL relies on deep learning.