bborja / mods_evaluation Goto Github PK
View Code? Open in Web Editor NEWMarine Obstacle Detection Benchmark - Evaluation and Visualization Scripts
Home Page: http://box.vicos.si/borja/viamaro/index.html
Marine Obstacle Detection Benchmark - Evaluation and Visualization Scripts
Home Page: http://box.vicos.si/borja/viamaro/index.html
I followed the steps and placed the results in the below folder structure.
I used the below YAML file for evaluation .
PATHS:
# Results output folder
RESULTS: 'D:/WACV2022/mods_new_output/'
# Path to where the dataset is stored
DATASET: 'D:/WACV2022/mods/'
# Path to where calibration files are stored
DATASET_CALIB: 'D:/WACV2022/mods/calibration'
# Path to where the segmentation predictions are stored
SEGMENTATIONS: 'D:/WACV2022/mods_new_output/'
SEGMENTATIONS:
# Prediciton dir structure:
# - False: pred_dir/method/sequence/frame.png
# - True: pred_dir/sequence/method/frame.png
SEQ_FIRST: False
# Specify which colors are used in segmentation maps
INPUT_COLORS:
- [0, 0, 0] #[247, 195, 37] # Obstacle
- [255, 0, 0] #[ 41, 167, 224] # Water
- [0, 255, 0] #[ 90, 75, 164] # Sky
VISUALIZATION:
# Display per-sequence evaluation progress
SEQUENCE_PROGRESS: True
overlap_percentages, overlap_percentages_d = run_evaluation_image(self.cfg, File "D:\WACV2022\mods_evaluation-master\mods_evaluation-master\modb_evaluation.py", line 407, in run_evaluation_image raise 'Missing number of something for sequence %s frame %d' % (seq_path_split[0], frame_number) TypeError: exceptions must derive from BaseException
Hi,
I would like to thank you for your massive contribution. I am only confused with the amount of data made available with MODS dataset.
The paper says: We present a new diverse maritime evaluation dataset containing approximately 81k stereo images synchronized with an on-board IMU, with over 60k objects annotated.
However, upon download I was only able to find 16350 frames, i.e. 8175 stereo images.
I am wondering, is this only a publicly available evaluation subset of it? Can the rest of the data be accessed?
In the paper you also state that you have merged the MODD1, MODD2 and SMD datasets into a single MODS training
dataset, but it sounds something different to me. Anyway, it results only in 24k images.
Hello,
First of all - thank you for building and sharing this awesome resource!
Next, a few clarifying questions:
Finally, a couple of issues I've run into when running the modb_evaluation.py
evaluation script.
i. I get a bunch of warnings such as this one
WARN:[email protected]] global /io/opencv/modules/imgcodecs/src/loadsave.cpp (239) findDecoder imread_('/stru02-00116500-00117000/frames/ignore_mask.png'): can't open/read file: check file path/integrity
(Indeed the ignore_mask.png
files appear not to be included in the dataset I've downloaded.)
ii. The script terminates prematurely with the following error message:
Traceback (most recent call last):
File "/home/oliskir/computer-vision/mods/mods_evaluation/modb_evaluation.py", line 471, in <module>
run_evaluation()
File "/home/oliskir/computer-vision/mods/mods_evaluation/modb_evaluation.py", line 324, in run_evaluation
write_json_file(cfg.PATHS.RESULTS, args.method, evaluation_results)
File "/home/oliskir/computer-vision/mods/mods_evaluation/utils/utils.py", line 197, in write_json_file
with open(os.path.join(output_path, output_file_name), 'w') as json_file:
FileNotFoundError: [Errno 2] No such file or directory: '../mods_wasr/results_../wasr_augm0_plainmaster4.json'
Probably something to do with the automatic construction of the output folder path? I've attached my config file to give you an idea of what my directory structure looks like.
mods.txt
Thanks!
Oliver
A declarative, efficient, and flexible JavaScript library for building user interfaces.
๐ Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. ๐๐๐
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google โค๏ธ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.