Search

Segmentation is an instrumental task in medical image analysis. In addition to existing manual, semi-automatic and automatic segmentation models, deep learning has been the niftiest machine learning technique in current research interests. However, none of the models or technique can escape from the overdependence on training data and user intervention. As a result, the use of computer-aided and learning algorithms have reported lackluster robustness in the presence of high anatomical disparity. In recognition of this, we have proposed a binary seeds auto-generation model to reduce the reliance on manually crafted priori information in deep learning. Then, we computed the reproducibility of the proposed model against manual segmentation using normal and osteoarthritic knee magnetic resonance image. In normal knee image, mean agreements of the proposed model and manual segmentation were $0.94 \pm 0.022$ and $0.83 \pm 0.028$ respectively. In osteoarthritic knee image, mean agreements of the proposed model and manual segmentation were $\mathbf{0.92}\pm\mathbf{0.051}$ and $\mathbf{0.79}\pm \mathbf{0.073}$ respectively. Pair t test showed that our method has better accuracy than manual segmentation in both cases (normal: $\text{P}=1.03\times 10^{-9}$; osteoarthritic: $\text{P}=4.94\times 10^{-8}$). Therefore, we can conclude the model is robust to be implemented as part of deep learning based segmentation framework.

Boundary correction is an important part in medical image segmentation process. Typically, refinement is needed because current segmentation techniques are not robust enough to cope with great anatomical variation demonstrated in medical image. In this works, we have proposed a boundary refinement method based on scribbling method. For instance, the scribbling based refinement method is based on “freehand draw” function. Our qualitative result shows that the scribbling refinement technique is highly adaptive to different knee cartilage geometries, which represents a challenge in the study of osteoarthritis. Future works should focus on adding pertinent features to minimize the degree of user intervention during boundary refinement.

Manual segmentation has been widely accepted in medical image analysis because the user will assume absolute control over final output. Nonetheless, the method is infamous for being time consuming and sensitive to inter-observer variability. Semi-automated methods incorporate user knowledge to initialize computerized segmentation and thus, is a versatile approach to tackle real time segmentation problem. Interactive tool or Graphic User Interface is the most critical medium of communication between user and computer but this element has always been neglected. Intuitively, the design of efficient interactive tool will impose significant effect on segmentation efficiency. In this work, we have presented a mini review on the design element of graphic user interface in medical image segmentation to highlight the importance of this tool in ensuring desirable segmentation output.