Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Duygu Balpetek Külcü has a total of 17 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Overcoming resolution constraints in automated colony counting via a high-performance deep learning framework using SAHI

    by Sercan Külcü, Duygu Balpetek Külcü

    Lightweight models perform poorly in bacterial colony counting when high-resolution Petri dish images are downscaled to 640 × 640 pixels. This study addresses this issue using a tiled training and SAHI-based tiled inference pipeline. Full-resolution images are divided into overlapping 640 × 640 tiles (20% overlap), preserving native resolution and reducing colony density per tile. A public 24-class bacterial colony dataset was used to evaluate the proposed approach. The tiled strategy substantially increased the number of training samples while preserving local colony details. Three nano-scale models (YOLOv5n, YOLOv8n, YOLOv11n) were evaluated. The conventional resize method yielded only 44.9-66.3% mAP@0.5, whereas the tiled approach achieved 95.4-96.9% mAP@0.5 and 55.4-58.4% mAP@0.5:0.95. YOLOv11n with tiling achieved the highest mAP@0.5 while using only 2.6 million parameters. It reached 96.9% mAP@0.5 and 58.2% mAP@0.5:0.95, and the validation-set confusion matrix showed class-wise correct prediction rates above 95% in 23 of 24 classes. Whole-plate inference required less than 320 ms on a laptop RTX 3050 GPU. For small, densely packed objects such as bacterial colonies, tiled inference with moderate overlap dramatically outperforms both image resizing and recent architectural advances in YOLO. The proposed lightweight pipeline is suitable for routine laboratory deployment.

  • Bias Mitigation in Ensemble-Based Meat Freshness Classification Using Grad-CAM

    by Sercan Külcü, Duygu Balpetek Külcü

    Visual biases in deep learning models, such as focusing on packaging trays instead of meat texture, reduce the reliability of computer vision systems in food safety applications. This study proposes a Grad-CAM-guided bias mitigation framework for multiclass meat freshness classification that combines explainable AI with a lightweight hybrid ensemble design. A MiniCAM attention module is integrated into MobileNetV2 to redirect model focus toward meat-specific visual cues, and its features are fused with complementary embeddings extracted from Xception. The final decision is obtained by combining the predictions of MobileNetV2 with classical classifiers (SVM and XGBoost) using test-time augmentation and grid-optimized weighted ensembling. The proposed framework achieves 99.78% accuracy on the held-out test set and 99.66% ± 0.23 average accuracy under 5-fold cross-validation, while maintaining real-time efficiency (4.3M parameters, 16.5 MB model size, and 825.1 FPS on a single GPU), and effectively suppresses non-informative background elements (e.g., packaging trays) as confirmed by Grad-CAM visualizations. These results demonstrate that integrating explainable bias mitigation with lightweight ensemble learning enables reliable and deployable meat freshness assessment for real-world food safety inspection.

  • Open Access

    Formulation and quality assessment of wheat germ and date powder-enriched biscuits: Rheological, textural, and microstructural properties

    by Duygu Balpetek Külcü, Elif Turabi Yolaçaner

    This study examines the incorporation of wheat germ (WG) and date powder (DP) into biscuit formulations to enhance their nutritional, rheological, and sensory properties. WG partially replaced wheat flour, while DP substituted sugar at different levels. Rheological analyses showed that WG increased dough strength and water absorption but lowered extensibility, whereas DP improved moisture retention and moderated the structural effects of WG. Both ingredients enhanced the nutritional profile by increasing dietary fiber, protein, and mineral contents. Texture and color analyses indicated that DP led to darker coloration due to Maillard reactions, while WG contributed to greater hardness. Moderate levels of both ingredients improved the biscuit’s internal structure and sensory acceptability. The optimized formulation provided a desirable balance of texture, flavor, and nutritional improvement. Overall, this study presents an approach to devel-oping functional bakery products by combining cereal by-products with natural sweeteners, supporting sus-tainable and health-focused food design.

  • Open Access

    Gıda Sektöründe Endüstri 4.0 ve Gıda Güvenliği

    by Hasret Türkoğlu, Duygu Balpetek Külcü

    Tarihin başlangıcından beri insanların yaşamlarını sürdürebilmeleri için en temel gereksinimleri gıda ürünleri (yiyecek-içecek) olmuştur. Artan insan nüfusu ile gıda ürünlerine olan ihtiyaç da artmış ve gelişen teknolojiyle birlikte, gıda sektöründe gerçekleştirilen seri üretimlerle bu ihtiyaçlar karşılanmaya çalışılmıştır. Ancak küreselleşmenin de etkisiyle artan kitlesel gıda ürünlerinin güvenliği sorunu ortaya çıkmış ve bu da toplumları gıda güvenliği ve izlenebilirliği politikaları ve izlenebilirlik sistemleri oluşturmaya yönlendirmiştir. Bu derlemede 18. yy’da buharlı makinelerin kullanımıyla ortaya çıkan 1. Endüstri Devrimi ile başlayıp günümüzde endüstriyel süreçlerde makinelerin insan gücüne göre etkin bir rol almaya başladığı 4. Endüstri Devrimi (Endüstri 4.0) dahil olmak üzere endüstriyel devrim aşamaları ele alınmış ve bunların gıda sektöründeki etkileri incelenmiştir. Endüstri 4.0’ün temel prensipleri ve uygulamalarının gıdada izlenebilirlikte kullanımı ile ilgili çalışmalar araştırılarak tartışılmıştır.


  • Open Access

    Determining Some Quality Characteristics of Vegan Tarhana Added with Red Beet (Beta Vulgaris Var. Cruenta) Powder

    by Sena Soğuksulu, Duygu Balpetek Külcü

    The present study aims to increase the number of tarhana versions, which have traditionally had a place in our culture, in order to increase the range of vegan products. In the present study, it was aimed to diversify the versions of tarhana, which has had a place in our traditional culture. For this purpose sauced and sauce-free vegan tarhana (yogurt-free) samples added with red beet (Beta vulgaris var. Cruenta) powder at different concentrations (control, 0.5%, 1%, 1.5%, 2%, 2.5%, 3%) samples were produced. Several physicochemical, microbiological, and sensorial properties of tarhanas produced were examined. Vegan tarhana with tomato sauce with 1.0% red beet powder had the highest score for consistency, taste, aroma and general acceptability. The highest score of color and odor properties was determined in vegan tarhana with tomato sauce added with 2.0% red beet powder. Ash, total acidity, protein, water-holding capacity, total dietary fiber, total phenolic content, and antioxidant capacity increased in parallel with red beet powder concentration. It was found that microbiological characteristics met the standards for tarhana, that the total number of mesophilic aerobic bacteria was higher than that of total yeast-mold. Staphylococcus aureus and coliform-group bacteria could not be detected in vegan tarhanas with red beet powder. In conclusion, it can be stated that all the vegan tarhanas with red beet powder samples had equivalent physicochemical and bioactive characteristics in comparison to the tarhana samples examined in the literature.


  • Investigation of physicochemical, microbiological, textural, and sensory properties of set‐type yogurt with Mentha pulegium L. (pennyroyal) powder

    by Duygu Balpetek Külcü, Emine Büşra Koşgin, Ömer Faruk Çelik, Elif Turabi Yolacaner

    In this study, five types of yogurt samples were produced with different concentrations (0% (control group), 0.05%, 0.10%, 0.15%, 0.20%) of Mentha pulegium L. (pennyroyal) powder (MPP) and their physicochemical, textural, sensory, and microbiological properties were investigated on the 1st, 7th, and 14th days of storage. Increasing concentrations of MPP affected the L * and a * values of yogurt considerably. 0.05% and 0.15% MPP addition improved the textural properties of yogurt samples. The higher the MPP concentration the lower the lactic acid bacteria counts were however, the numbers were still over the minimum requirement for yogurt. Fibrous and filamentous structures in yogurt samples with MPP were monitored by scanning electron microscopy. Consequently, it can be stated that MPP can be successfully added to yogurt. Furthermore, considering sensory evaluations, it would be plausible to say that yogurt samples containing MPP up to 0.10% may have market potential in the future.


  • Open Access

    Determination of the Microflora of Meatless Cig Kofte in The Black Sea Region of Turkey

    by Duygu Balpetek Külcü, Sena Soğuksulu, Yunus Çakan, Çağla Ergen

    Objective Meatless Cig Kofte (MCK) is a traditional product consumed frequently in Turkey. In recent years, meatless raw meatball shops have been opened and offered to the public. It is important for public health because it is a risky food for microbial contamination. This study was conducted to investigate some microbiological quality of MCK sold in Black Sea region (Trabzon, Giresun, Ordu, Samsun) in Turkey. Methods: In this study, a total of 120 MCK were sampled from different points of sales and examined for the microbiological quality of MCK in the Black Sea Region (Trabzon, Giresun, Ordu and Samsun). Data analysis of the study was conducted using SPSS 24 package program. Statistical significance was evaluated at p<0.05. Results: According to the research results, the highest average TMAB (Total Mesophilic Aerobic Bacteria) value of MCK samples was determined in Giresun province (6.04 log cfu/g) and the lowest average value was found in Samsun province. While the average TPAB (Total Psychophilic Aerobic Bacteria) was determined as 1.82 log cfu/g in MCK samples collected from Samsun province, it could not be detected in Trabzon, Ordu and Giresun provinces. The average S. aureus value of MCK samples was determined in the highest province of Ordu (4.22 log cfu/g) and the lowest in Trabzon province (1.94 log cfu/g). Conclusion: It was concluded that microbial quality was insufficient and hygiene rules were not taken into consideration in the provinces that were included in this study conducted in the Black Sea Region (Trabzon-Giresun-Ordu-Samsun).


  • Open Access

    Consciousness Level Determination of Red Meat Consumption of Pregnant Women, Giresun/Turkey Province

    by Duygu Balpetek Külcü, Özge Cağcağ Yolcu

    In this study, meat consumption and the nutritional status of pregnant women living in Giresun province of Turkey were investigated. The study was carried out at Giresun University, the Maternity and Pediatrics Training and Research Hospital between February-March 2019. The study group consisted of 218 pregnant women aged 18 to 49 years, who benefited from polyclinic services and agreed to participate in the survey. A questionnaire consisting of 12 questions was used, and the findings were evaluated using SPSS to generate descriptive statistics and t-tests or ANOVAs. The majority of participants stated that red meat is beneficial during pregnancy (71.1%) and this benefit is due to its nutrients (69.1%). Educational level was positively related to a balanced diet during pregnancy. Although the pregnant women in this study have knowledge about nutrition during pregnancy, nutrition programs can be further strengthened to increase their level of knowledge.