KMITL

Permanent URI for this communityhttps://dspace.kmitl.ac.th/handle/123456789/1

Browse

Search Results

Now showing 1 - 7 of 7

Some of the metrics are blocked by your
consent settings
Convolutional neural networks based focal loss for class imbalance problem: a case study of canine red blood cells morphology classification
(2023-11-01)
Pasupa, Kitsuchart
;
Vatathanavaro, Supawit
;
Tungjitnob, Suchat
Morphologies of red blood cells are normally interpreted by a pathologist. It is time-consuming and laborious. Furthermore, a misclassified red blood cell morphology will lead to false disease diagnosis and improper treatment. Thus, a decent pathologist must truly be an expert in classifying red blood cell morphology. In the past decade, many approaches have been proposed for classifying human red blood cell morphology. However, those approaches have not addressed the class imbalance problem in classification. A class imbalance problem—a problem where the numbers of samples in classes are very different—is one of the problems that can lead to a biased model towards the majority class. Due to the rarity of every type of abnormal blood cell morphology, the data from the collection process are usually imbalanced. In this study, we aimed to solve this problem specifically for classification of dog red blood cell morphology by using a Convolutional Neural Network (CNN)—a well-known deep learning technique—in conjunction with a focal loss function, adept at handling class imbalance problem. The proposed technique was conducted on a well-designed framework: two different CNNs were used to verify the effectiveness of the focal loss function and the optimal hyperparameters were determined by fivefold cross-validation. The experimental results show that both CNNs models augmented with the focal loss function achieved higher F<inf>1</inf> -scores, compared to the models augmented with a conventional cross-entropy loss function that does not address class imbalance problem. In other words, the focal loss function truly enabled the CNNs models to be less biased towards the majority class than the cross-entropy did in the classification task of imbalanced dog red blood cell data.
Some of the metrics are blocked by your
consent settings
CowXNet: An automated cow estrus detection system
(2023-01-01)
Lodkaew, Thanawat
;
Pasupa, Kitsuchart
;
Loo, Chu Kiong
Estrus detection is essential for dairy farms to take cows for artificial insemination promptly. Conventional approaches for detecting estrus cows use electronic devices attached to cows to gather data for software analysis. However, electronic devices can be costly and make a cow moody and uncomfortable while moving. In a common approach, observers detect estrus cows by observing their behaviors. However, continuous observation can easily lead to errors due to the observer's fatigue. Therefore, we designed CowXNet, an automatic estrus detection system for cows, to assist farmers to detect estrus cows. CowXNet requires only a camera attached in a pen and a computer to analyze recorded videos. CowXNet analyzes the estrus behaviors of each cow in a pen and helps farmers to identify estrus cows. To develop and evaluate CowXNet efficiently and effectively, we collected data from Chokchai Farm, the biggest dairy farm in Asia (14.65483∘N, 101.34853∘E). CowXNet has four modules: (i) cow detection uses YOLOv4 to detect cows in recorded videos; (ii) body part detection uses a convolutional neural network to estimate locations of body parts of detected cows; (iii) estrus behavior detection uses body part coordinates to extract a set of discriminative features, and a classification algorithm to detect estrus behaviors, and (iv) behavior analysis module displays estrus behavior for analysis purposes. We evaluated CowXNet for two instances: module-independent evaluation and end-to-end framework evaluation. Overall, CowXNet was promising; it correctly detected estrus behavior interval of cows 83% of cases.
Some of the metrics are blocked by your
consent settings
Evaluation of deep learning algorithms for semantic segmentation of car parts
(2022-10-01)
Pasupa, Kitsuchart
;
Kittiworapanya, Phongsathorn
;
Hongngern, Napasin
;
Woraratpanya, Kuntpong
Evaluation of car damages from an accident is one of the most important processes in the car insurance business. Currently, it still needs a manual examination of every basic part. It is expected that a smart device will be able to do this evaluation more efficiently in the future. In this study, we evaluated and compared five deep learning algorithms for semantic segmentation of car parts. The baseline reference algorithm was Mask R-CNN, and the other algorithms were HTC, CBNet, PANet, and GCNet. Runs of instance segmentation were conducted with those five algorithms. HTC with ResNet-50 was the best algorithm for instance segmentation on various kinds of cars such as sedans, trucks, and SUVs. It achieved a mean average precision at 55.2 on our original data set, that assigned different labels to the left and right sides and 59.1 when a single label was assigned to both sides. In addition, the models from every algorithm were tested for robustness, by running them on images of parts, in a real environment with various weather conditions, including snow, frost, fog and various lighting conditions. GCNet was the most robust; it achieved a mean performance under corruption, mPC = 35.2, and a relative degradation of performance on corrupted data, compared to clean data (rPC), of 64.4%, when left and right sides were assigned different labels, and mPC = 38.1 and rPC = 69.6 % when left- and right-side parts were considered the same part. The findings from this study may directly benefit developers of automated car damage evaluation system in their quest for the best design.
Some of the metrics are blocked by your
consent settings
Hybrid Deep Learning Models for Thai Sentiment Analysis
(2022-01-01)
Pasupa, Kitsuchart
;
Seneewong Na Ayutthaya, Thititorn
Many people use social media in their daily life for entertainment, business, personal communication, and catching up with friends. In social media marketing, sentiment analysis is one of the most popular research topics because it can be employed to perform brand or market research monitoring and to keep an eye on the competitors. Machine learning algorithms have been utilized to carry out the task. In addition, sentiment analysis is essential in cognitive computing. Currently, there are still a limited number of Thai sentiment analysis research. This paper proposes a framework for sentiment analysis in Thai along with Thai-SenticNet5 corpus. The framework employs different types of features, namely, word embedding, part-of-speech, sentic features, and all combinations of these features. Furthermore, we fused deep learning algorithms—convolutional neural network (CNN) and bidirectional long short-term memory (BLSTM)—in different ways and compare it to several other fused combinations. Three datasets in Thai were used in this work: ThaiTales, ThaiEconTwitter, and Wisesight datasets. The experimental results show that combining all three features and fusing deep learning algorithms were able to improve overall performance. The best hybrid deep learning was BLSTM-CNN that achieved F<inf>1</inf>-scores of 0.7436, 0.7707, and 0.5521, on ThaiTales, ThaiEconTwitter, and Wisesight datasets, respectively. According to the experimental results, we conclude that feature combination and hybrid deep learning algorithms can improve the overall performances.
Some of the metrics are blocked by your
consent settings
Deep Learning-Based Early Detection and Avoidance of Traffic Congestion in Software-Defined Networks
(2022-01-01)
Prabhavat, Sumet
;
Thongthavorn, Thananop
;
Pasupa, Kitsuchart
Software-defined Networking (SDN) provides an easy way to monitor network and traffic conditions by employing software-based controllers to communicate with the hardware directly. It provides helpful information that enables efficient routing decisions. This research study attempted to use deep learning techniques - Long Short-term Memory, Bidirectional Long Short-term Memory, and Gated Recurrent Unit - to predict network traffic to allow the controller to early detect congestion. The traffic flow in a network link that will likely be congested will be rerouted to a new path with the largest available bandwidth. Various scenarios were simulated to evaluate our deep learning-based SDN controller (Ryu controller platform). The results show that our proposed deep learning-based SDN controller outperformed the traditional load balancing technique.
Some of the metrics are blocked by your
consent settings
Improved Identification of Imbalanced Multiple Annotation Intent Labels with a Hybrid BLSTM and CNN Model and Hybrid Loss Function
(2021-01-01)
Vatathanavaro, Supawit
;
Pasupa, Kitsuchart
;
Sirirattanajakarin, Sorratat
;
Suntisrivaraporn, Boontawee
Payment or fund transfer transactions can be annotated by users when they are made through a mobile banking app, for example, SCB Easy app—a mobile banking app by Siam Commercial Bank—allows users to annotate transactions with 40 character texts. The AI2 framework was used to identify user intentions with the transactions, so that the bank can offer the right product to the right customer at the right time. The framework employed Long Short-Term Memory (LSTM). Commonly, one annotated sample can be interpreted as representing multiple intents, thus we had a multiple label classification problem. However, the original model did not consider the class imbalance, that caused the model to bias toward the majority class. We introduced a new hybrid Bidirectional LSTM and Convolutional Neural Network model in conjunction with a new hybrid loss function to tackle the imbalance. Our model with hybrid loss function performed better than the AI2 framework with a 4.5% improvement in F<inf>1</inf> -score. Moreover, our hybrid loss function enabled the model to classify minority classes better, when the imbalance ratio became higher, compared with a conventional cross-entropy loss function. In other words, our hybrid loss function made the model to be more efficient in real-world multiple label imbalance problem.
Some of the metrics are blocked by your
consent settings
Thai sentiment analysis with deep learning techniques: A comparative study based on word embedding, POS-tag, and sentic features
(2019-10-01)
Pasupa, Kitsuchart
;
Seneewong Na Ayutthaya, Thititorn
A smart city connects physical, information technology, social, and business infrastructures together to leverage their collective intelligence. Feedback drives improvements in service, city development, and quality of life in the city. Therefore, sentiment analysis in real-time of opinions expressed in text form by residents in the city is absolutely necessary. Nowadays, machine learning is widely applied to sentiment analysis of decisions in business, especially deep learning. In this experiment, we evaluated and compared the performances of several conventional deep learning models: Convolutional Neural Network (CNN), Long Short-Term Memory (LSTM), and Bidirectional LSTM (Bi-LSTM), in sentiment analysis of Thai children tales. In several previous studies, many features have been used in all of the models mentioned, features such as word embedding that helps a model to understand the semantics of each word, POS-tag that helps a model to understand the grammatical function of words, and sentic that helps a model to understand the emotion of words. Some combinations of these features have also been used. The results of this experiment show that the CNN model that used all three features gave the best result of 0.817 F1-score at p < 0.01, which was significantly better than all other models.

KMITL

Browse

Filters

Settings

Sort By

Results per page

Search Results