Figure 3 - available via license: Creative Commons Attribution 4.0 International
Content may be subject to copyright.
Source publication
In this study, we propose a multi-scale ensemble learning method for thermal image enhancement in different image scale conditions based on convolutional neural networks. Incorporating the multiple scales of thermal images has been a tricky task so that methods have been individually trained and evaluated for each scale. However, this leads to the...
Context in source publication
Context 1
... this module, the feature maps from the feature extraction module are non-linearly mapped (see Figure 3). The first layer, namely shrinking, and the last layer, namely expansion, reduce and expand the number of channels of the feature map by a convolutional layer with a 1 × 1 kernel, respectively. ...
Citations
... Wang et al. [23] proved that the performance of a random forest which is an ensemble model of decision trees is superior to the SVM model for the classification of sacral regions in skin thermal images for the detection of pressure injury. Ban et al. [24] proposed an ensemble model of multi-scale convolution networks for thermal image enhancement. L. S. Garia et al. [25] presented an ensemble model of various pre-trained CNN models for the classification of breast thermograms for cancer detection. ...
In the field of medicine, thermal image processing and analysis play a significant role in the diagnosis, monitoring, and treatment of diseases. For example, during the last decade, several studies have been performed based on thermal image processing for ocular disease diagnosis. This research proposes a unique approach for the classification of subgroups of two retinal vascular diseases, namely diabetic eye disease and age-related macular degeneration (AMD). The class imbalance problem is a well-known issue when working with medical data, where one class is significantly less represented than another class in the dataset. To deal with the class imbalance issue, an ensemble decision tree classifier with a random under-sampling and adaptive boosting (RUSBoost) technique is proposed. The performance of the proposed classifier is compared with various traditional machine learning-based classifiers. Experimental results show that the proposed ensemble tree outperforms other classifiers through high accuracy, [Formula: see text]-score, and Mathews correlation coefficient (MCC) values in classifying diabetic eye diseases and AMD diseases. The proposed ensemble decision tree distinguishes dry AMD and wet AMD over healthy controls with 95% average accuracy. Also, it classifies diabetic retinopathy (DR) with diabetic macular edema (DME) and DR without DME with 94% average accuracy. The classifier could distinguish dry and wet AMD which did not work around in temperature analysis on the manual temperature measurement. The performance of the automated classification model is on par with the performance of the temperature analysis of OST for DME and DR without DME.
... Thermal imaging converts the temperature distribution emitted by objects into a visible image. In the paper entitled 'Multi-Scale Ensemble Learning for Thermal Image Enhancement', Yuseok Ban, and Kyungjae Lee [14] proposed a multi-scale ensemble learning method in different image scale conditions, which has a novel parallel architecture leveraging the confidence maps of multiple scales. ...
Intelligent image and optical information processing have paved the way for the recent epoch of new intelligence and information [...]
... For example, in the studies on image quality metric, many efforts have been made to find appropriate metrics for thermal images [22][23][24]. Further, in the studies on image enhancement, many research proposals have been made to develop methods specialized for thermal images to solve problems such as low signal-to-noise ratio (SNR), halo effect, blurring, and low dynamic range compared to visible images [25][26][27]. ...
Various types of motion blur are frequently observed in the images captured by sensors based on thermal and photon detectors. The difference in mechanisms between thermal and photon detectors directly results in different patterns of motion blur. Motivated by this observation, we propose a novel method to synthesize blurry images from sharp images by analyzing the mechanisms of the thermal detector. Further, we propose a novel blur kernel rendering method, which combines our proposed motion blur model with the inertial sensor in the thermal image domain. The accuracy of the blur kernel rendering method is evaluated by the task of thermal image deblurring. We construct a synthetic blurry image dataset based on acquired thermal images using an infrared camera for evaluation. This dataset is the first blurry thermal image dataset with ground-truth images in the thermal image domain. Qualitative and quantitative experiments are extensively carried out on our dataset, which show that our proposed method outperforms state-of-the-art methods.
