Dataset Information

Malaria parasite detection in thick blood smear microscopic images using modified YOLOV3 and YOLOV4 models.

ABSTRACT:

Background

Manual microscopic examination of Leishman/Giemsa stained thin and thick blood smear is still the "gold standard" for malaria diagnosis. One of the drawbacks of this method is that its accuracy, consistency, and diagnosis speed depend on microscopists' diagnostic and technical skills. It is difficult to get highly skilled microscopists in remote areas of developing countries. To alleviate this problem, in this paper, we propose to investigate state-of-the-art one-stage and two-stage object detection algorithms for automated malaria parasite screening from microscopic image of thick blood slides.

Results

YOLOV3 and YOLOV4 models, which are state-of-the-art object detectors in accuracy and speed, are not optimized for detecting small objects such as malaria parasites in microscopic images. We modify these models by increasing feature scale and adding more detection layers to enhance their capability of detecting small objects without notably decreasing detection speed. We propose one modified YOLOV4 model, called YOLOV4-MOD and two modified models of YOLOV3, which are called YOLOV3-MOD1 and YOLOV3-MOD2. Besides, new anchor box sizes are generated using K-means clustering algorithm to exploit the potential of these models in small object detection. The performance of the modified YOLOV3 and YOLOV4 models were evaluated on a publicly available malaria dataset. These models have achieved state-of-the-art accuracy by exceeding performance of their original versions, Faster R-CNN, and SSD in terms of mean average precision (mAP), recall, precision, F1 score, and average IOU. YOLOV4-MOD has achieved the best detection accuracy among all the other models with a mAP of 96.32%. YOLOV3-MOD2 and YOLOV3-MOD1 have achieved mAP of 96.14% and 95.46%, respectively.

Conclusions

The experimental results of this study demonstrate that performance of modified YOLOV3 and YOLOV4 models are highly promising for detecting malaria parasites from images captured by a smartphone camera over the microscope eyepiece. The proposed system is suitable for deployment in low-resource setting areas.

SUBMITTER: Abdurahman F

PROVIDER: S-EPMC7938584 | biostudies-literature | 2021 Mar

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Malaria parasite detection in thick blood smear microscopic images using modified YOLOV3 and YOLOV4 models.

Abdurahman Fetulhak F Fante Kinde Anlay KA Aliy Mohammed M

BMC bioinformatics 20210308 1

<h4>Background</h4>Manual microscopic examination of Leishman/Giemsa stained thin and thick blood smear is still the "gold standard" for malaria diagnosis. One of the drawbacks of this method is that its accuracy, consistency, and diagnosis speed depend on microscopists' diagnostic and technical skills. It is difficult to get highly skilled microscopists in remote areas of developing countries. To alleviate this problem, in this paper, we propose to investigate state-of-the-art one-stage and two ...[more]

PMID: 33685401

Similar Datasets

Project description:BackgroundDespite many technological advances for malaria parasite detection (e.g. high resolution image acquisition), microscopic reading of thick blood smear (TBS) remains the gold standard. Even though available in low technology environment, the microscopy of TBS is slow and time consuming. Moreover microscopy may induce errors at many levels and has no quality control.MethodsA electronic extension of the mechanical tally counter is proposed. In addition to the counting process it includes the process of counting itself that relies on the time elapsed between two successive pressures of the counting button leading to a timed tally counter (TTC). The microscopist performs the reading with the specific instruction starting by counting, in each high power fields, leucocytes first and then parasites. The time-stamp of all pressures of counting buttons are recorded along with the nature of the count. The data are recorded internally in CSV format and are exportable. The detection of HPFs locations and leukocyte/parasite counts per HPFs is performed through a hidden semi-Markov model (with outliers) allowing both to take into account the known distribution of leukocyte per HPFs (using a negative binomial distribution) and the pauses and hesitation of the microscopist during the reading. Parameters are estimated via the expectation-maximization algorithm. Hyper-parameters are calibrated using expert annotations. Forward/backward recursions are used to obtain the HPFs locations.ResultsThis approach provides richer data at no extra cost. It has been demonstrated that the method can derive parasites per HPF, leukocytes per HPF, and parasite/leukocyte ratio with robust non-parametric confidence intervals. Moreover a direct digital data entry leads to a less expensive process and decreased time-consuming and error-prone manual data entry. Lastly the TTC allows detecting possible protocol break during reading and prevents the risk of fraud.Discussion and conclusionIntroducing a programmed digital device in the data acquisition of TBS reading gives the opportunity to develop easily new (possible adaptive) reading protocols that will be easily followed by the reader since they will be embedded directly in the device. With the TTC the reader only has to read HPFs, counting leukocytes first and parasites second, and the counter will beep when the protocol is completed.

Project description:BackgroundAmbient temperature is an important determinant of malaria transmission and suitability, affecting the life-cycle of the Plasmodium parasite and Anopheles vector. Early models predicted a thermal malaria transmission optimum of 31 °C, later revised to 25 °C using experimental data from mosquito and parasite biology. However, the link between ambient temperature and human malaria incidence remains poorly resolved.MethodsTo evaluate the relationship between ambient temperature and malaria risk, 5833 febrile children (<18 years-old) with an acute, non-localizing febrile illness were enrolled from four heterogenous outpatient clinic sites in Kenya (Chulaimbo, Kisumu, Msambweni and Ukunda). Thick and thin blood smears were evaluated for the presence of malaria parasites. Daily temperature estimates were obtained from land logger data, and rainfall from National Oceanic and Atmospheric Administration (NOAA)'s Africa Rainfall Climatology (ARC) data. Thirty-day mean temperature and 30-day cumulative rainfall were estimated and each lagged by 30 days, relative to the febrile visit. A generalized linear mixed model was used to assess relationships between malaria smear positivity and predictors including temperature, rainfall, age, sex, mosquito exposure and socioeconomic status.ResultsMalaria smear positivity varied between 42-83% across four clinic sites in western and coastal Kenya, with highest smear positivity in the rural, western site. The temperature ranges were cooler in the western sites and warmer in the coastal sites. In multivariate analysis controlling for socioeconomic status, age, sex, rainfall and bednet use, malaria smear positivity peaked near 25 °C at all four sites, as predicted a priori from an ecological model.ConclusionsThis study provides direct field evidence of a unimodal relationship between ambient temperature and human malaria incidence with a peak in malaria transmission occurring at lower temperatures than previously recognized clinically. This nonlinear relationship with an intermediate optimal temperature implies that future climate warming could expand malaria incidence in cooler, highland regions while decreasing incidence in already warm regions with average temperatures above 25 °C. These findings support efforts to further understand the nonlinear association between ambient temperature and vector-borne diseases to better allocate resources and respond to disease threats in a future, warmer world.

Dataset Information

Malaria parasite detection in thick blood smear microscopic images using modified YOLOV3 and YOLOV4 models.

Background

Results

Conclusions

Publications

Malaria parasite detection in thick blood smear microscopic images using modified YOLOV3 and YOLOV4 models.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets