Project's information

Project's title Research and Develop methods for automatic detection of morphometric landmarks on insect wing images.
Project’s code VAST01.01/19-20
Research hosting institution Hanoi University of Science & Technology
Project leader’s name Nguyen Hoang Ha
Project duration 01/01/2019 - 31/12/2021
Project’s budget 600 million VND
Classify Fair
Main results

Theoretical results: 01 article accepted for publication on the Ecological Informatics journal (indexed in SCI/SCIE, ranked Q1 in May 2022) and 02 Master students supported.
Applied results: 01 open source and free prototype software titled iMorph for automatic landmark detection in insect wing images and fine tuning landmark position.

Novelty and actuality and scientific meaningfulness of the results

The novel contributions of this project include 02 methods for detection of morphometric landmark in insect wing images. The first one is a classical method based upon handcrafted feature extraction and key-point matching techniques. The second method is based on deep learning method that in turn consists of two approach: using existing object detection methods, and building a two-phrase regression deep learning process for coordinate prediction.

Products of the project

- Scientific papers in referred journals (list):
Hoang Ha Nguyen, Bich Hai Ho, Hien Phuong Lai, Hoang Tung Tran, Anne Laure Banuls, Jorian Prudhomme, and Huu Ton Le, A lightweight keypoint matching framework for insect wing morphometric landmark detection, Ecological Informatics, ISSN 1574-9541, accepted May 2022.
- Patents (list):
- Technological products (describe in details: technical characteristics, place):
o    Open source and cross platform (Windows, MacOS, Linux) software titled iMorph  allowing ecologists annotate the morphometric landmarks in insect wing images. It is available at: https://github.com/hausth/WingLanmarkPredictor
o    Dữ liệu: 5 public datasets have been collected then gathered at: https://github.com/ha-usth/InsectWingLandmarkDatasets there in the Droso-big dataset has been annotated manually according to the same way with Droso-small (Vandaele, R. et al. Landmark detection in 2D bioimages for geometric morphometrics: a multi-resolution tree-based approach. Sci. Reports 8, 538, DOI: 10.1038/s41598-017-18993-5 (2018)
- Other products (if applicable):    
+  Supervision of the master thesis of Huỳnh Vinh Nam titled “Automatic detection of morphometric landmarks in insect wings using deep learning methods”, University of Science and Technology of Hanoi - VAST, defended in 7/2021.
+  Co-supervision of the master thesis of Tô Hồng Quân titled “Research some characteristics of digital image interpolation using some morphological operations to improve image quality”, Graduate University of Science and Technology - VAST, defended in 11/2021

Research region

The iMorph open source software is applicable at the Institute of Ecology and Biological Resource - VAST.

Images of project
1671511069083-180.png