1st International and 10th National Iranian Conference on Bioinformatics
Home Accepted Papers
Machine Vision Based Conic Estimation: An Effective Tool for Reinforcement of Deep Learning in Measuring Fetal Head Circumference
Paper ID : 1363-ICB10
Authors:
Amir Saniyan1, Seyed Vahab Shojaedini *2
1Department of Computer Engineering, Qazvin Branch, Islamic Azad University, Qazvin, Iran.
2Department of Biomedical Engineering, Iranian Research Organization for Science and Technology, Tehran, Iran.
Abstract:
Abstract
The artificial intelligence is considered as an effective way to estimate the border around the fetal head based on ultrasonic images [1]. This approach has been remained as an open problem because of existing some challenges such as low contrast of the image under various conditions and its low signal to noise ratio [2].

Targets
In this article the machine vision based, conic estimation scheme is utilized as supplementary processing of a VNET deep network in order to promote the performance of artificial intelligenece in measuring fetal head parameters [3]. Firstly, a cost function is constructed by using a family of curves, each of them may be a candidate for the fetal border. Then the objective function is minimized by using optimization framework in order to obtain the best fitted curve.

Methodology
The testbed for evaluation of the proposed idea was prepared by using the software Python with Keras on a computer with a NVIDIA 2080 TI GPU with 32 GB RAM. The tests were performed on HC18 challenge dataset provided by Heuvel and collaborators [4] which includes 999 ultrasound images which have been divided into three categories of training, validation and test subsets, with ratios of 60%, 15% and 25% among the total images respectively.

Founded
The obtained results showed that proposed scheme may lead to significant promotion in measuring the parameters of fetal head. Therefore our method resulted in the accuracy, precision dice and jaccard parameters equal to 97.6%, 94.3%, 95.4% and 91.7% respectively. These results show at least 3.6% improvement over the same methods which did not use from geometric concept.

Conclusions
Improving the estimation parameters show effectiveness of our idea based on reinforcement of artificial intelligence tools in measuring fetal head circumference by considering the geometric nature of the problem.
Keywords:
Fetal head circumference; Deep learning; Conic fitting; Geometric properties.
Status : Paper Accepted (Poster Presentation)