Bread
Description
Project Title: Good and Bad Classification of Bread (Triticum Aestivum) Using Realme 11x 5G Mobile Camera Description: The project, titled "Good and Bad Classification of Bread (Triticum Aestivum)," focuses on developing a machine learning system to automatically classify bread into two categories: good (fresh and high-quality) and bad (stale or low-quality). This classification task is based on a dataset of over 500 images, equally divided between good and bad bread samples. The images were captured using the Realme 11x 5G mobile camera, with bread samples photographed against a black background in daylight conditions, ensuring clarity and uniformity in the captured data. Dataset Composition: Good Samples (Fresh Bread): More than 250 images in the dataset depict good quality bread. These bread samples display characteristics of freshness, including an even golden-brown crust, soft and uniform texture, and a lack of defects such as burns or cracks. These images form the positive class, representing the standard of high-quality, freshly baked bread. Bad Samples (Stale or Poor-Quality Bread): The dataset also includes over 250 images of bad quality bread. These samples may show signs of staleness (hard or dry texture), uneven baking (burns, undercooked sections), or spoilage (mold, discoloration). These bad examples form the negative class and are crucial for training the model to recognize undesirable qualities in bread. Data Collection Setup: The bread images were captured using a Realme 11x 5G mobile camera, which offers high-resolution imaging, ensuring that the fine details of the bread samples, such as texture, color, and crust appearance, are clearly visible. A black background was used to create contrast with the bread, making it easier to highlight the defining features of the samples. Natural daylight conditions were chosen for consistent and even lighting across all images, reducing the influence of shadows or artificial light variations. Image Characteristics: The dataset features a wide variety of bread types and conditions. The good and bad bread samples differ in size, shape, texture, and quality, providing a broad range of visual examples for the classification model. This variability is crucial for ensuring the model can generalize well and classify different types of bread, irrespective of their size, shape, or baking method. Data Annotation: Each image is labeled as either "good" or "bad" based on the visual quality of the bread. Expert observation was used to categorize the bread samples, ensuring that the labels are accurate and reliable for training the classification model. These annotations serve as the ground truth for supervised learning, enabling the model to distinguish between fresh and stale or spoiled bread effectively. Data Preprocessing: The dataset undergoes several preprocessing steps to ensure the machine learning model receives optimal inputs: