MIT Licensehttps://opensource.org/licenses/MIT
License information was derived automatically
This dataset is designed to solve the task of detecting cows in aerial images, aiding in applications such as livestock counting and monitoring. The dataset consists of a single class:
Cows in this dataset appear as distinct shapes on open landscapes. From an aerial view, cows may seem small but are recognizable by their characteristic body shape and size, typically larger than other features in the landscape such as rocks or bushes.
Bounding Rules: Draw a bounding box around each cow visible in the image. Each box should encompass the entire cow's visible body, including any extremities like tails or legs if visible.
Object Appearance: Cows often appear as elongated shapes from above. They are typically situated in open areas and may cast a shadow, which should not be included in the annotation.
Occlusion: If a cow is partially occluded by terrain, include the visible portions in the bounding box as accurately as possible. Do not annotate if less than 20% of the cow is visible.
Size Constraints: Do not label objects smaller than 10 pixels in either dimension to ensure that the object is a cow.
Clarity and Identification: Only annotate if the object clearly resembles a cow. Do not include any objects that might be confused with other elements like rocks or terrain features unless you can clearly distinguish the cow-like form.
While annotating, ensure each cow is bounded separately, avoiding overlaps in annotation with other cows or labels.
MIT Licensehttps://opensource.org/licenses/MIT
License information was derived automatically
This dataset is for object detection, focusing on identifying and annotating planes and ships. The primary goal is to detect these objects in various scenes to assist in automated systems for recognition and tracking.
Planes are generally characterized by their distinct wing structures, elongated fuselage, and sometimes visible tails. They often have a central body from which wings extend horizontally, forming a cross-like appearance from a top view.
Ships are discernible by their long, narrow hulls often accompanied by various structures on top like decks or containers. They usually appear elongated with linear or curved features along the sides, often giving a top-down view a rectangular shape.
MIT Licensehttps://opensource.org/licenses/MIT
License information was derived automatically
This dataset focuses on the detection and classification of various laboratory caps and tubes. The goal is to facilitate automated sorting and organization processes in laboratory settings. The dataset contains five classes:
The 15ml cap is a small, circular object that features ridged edges for grip. It fits snugly onto 15ml tubes, and its top surface is often smooth and flat.
The 15ml tube is a medium-sized, cylindrical object that may be either transparent or translucent. It is typically seen from a side view.
The 50ml cap is a large, circular object similar in shape to the 15ml cap but larger in diameter. It is used for sealing 50ml tubes and has similar ridged edges.
The 50ml tube is a larger cylindrical tube, often used for larger volumes of samples. It is usually seen in a vertical position.
The 5ml cap is the smallest among the caps and is used for sealing 5ml tubes. It is circular with a much smaller diameter compared to the other caps.
MIT Licensehttps://opensource.org/licenses/MIT
License information was derived automatically
This dataset focuses on annotating electronic components on printed circuit boards (PCBs). The goal is to identify and label specific components using bounding boxes for component recognition and classification. The classes in the dataset include various components such as capacitors, integrated circuits (ICs), LEDs, and resistors, along with their respective footprints.
A rectangular component with metallic ends, often mounted flat on the PCB.
Annotate the entire visible body of the capacitor, ensuring the bounding box encloses both metallic ends and the main body. Ignore any solder connections not integral to the component itself.
Patterned area on the PCB indicating where a capacitor should be placed, usually marked with pads or outlines.
Enclose the marked pads or outlines on the PCB dedicated for capacitor placement. Do not include adjacent traces unless they are part of the footprint.
The bottom view of an Integrated Circuit, showing pins and a dark, flat body.
Draw bounding boxes around the visible chip body and pins. Ensure all pins are included within the annotation, focusing solely on the component itself.
The outlined area on the PCB that interfaces with IC pins, typically a rectangular set of pads.
Outline the entire footprint marked on the PCB for an IC, ensuring that all pads designed for pin connections are within the bounding box.
The top view of an Integrated Circuit, characterized by a smooth, often branded surface with pins around its perimeter.
Enclose the entire top surface of the IC along with its visible pins. Ensure any labels or marks on the top surface are centered in the annotation.
The underside of a Light Emitting Diode, generally a small square or rectangular component with metallic contacts.
Draw bounding boxes including the entire body of the LED and its metallic contacts. Do not include any mounting or surrounding board elements.
The designated area on the PCB for an LED, typically featuring small pads for connections.
Outline the footprint area on the PCB, capturing all pads and marks meant for LED placement. Adjacent circuit elements are to be ignored unless part of the footprint.
The top view of an LED, often small with a slightly raised dome or flattened surface.
Capture the entire body of the LED in the bounding box, including any visible lens or dome shape. Ensure no surrounding components are included.
A generally rectangular component with metallic ends, often seen mounted horizontally on the board.
Include the complete body of the resistor and its metallic ends in the annotation. Avoid including excess solder or trace areas unless directly attached to the resistor body.
The marked area on a PCB for resistor placement, featuring two pads or outlined areas.
Enclose the marked zone for resistor placement without extending the box beyond the designated pads or outlines.
The top view of a resistor, usually showing a smoother body with possible label markings or color bands.
Annotate the whole top-facing surface of the resistor, ensuring any visible markings are within the bounding area. Exclude unnecessary board elements not part of the resistor.
Not seeing a result you expected?
Learn how you can add new datasets to our index.
MIT Licensehttps://opensource.org/licenses/MIT
License information was derived automatically
This dataset is designed to solve the task of detecting cows in aerial images, aiding in applications such as livestock counting and monitoring. The dataset consists of a single class:
Cows in this dataset appear as distinct shapes on open landscapes. From an aerial view, cows may seem small but are recognizable by their characteristic body shape and size, typically larger than other features in the landscape such as rocks or bushes.
Bounding Rules: Draw a bounding box around each cow visible in the image. Each box should encompass the entire cow's visible body, including any extremities like tails or legs if visible.
Object Appearance: Cows often appear as elongated shapes from above. They are typically situated in open areas and may cast a shadow, which should not be included in the annotation.
Occlusion: If a cow is partially occluded by terrain, include the visible portions in the bounding box as accurately as possible. Do not annotate if less than 20% of the cow is visible.
Size Constraints: Do not label objects smaller than 10 pixels in either dimension to ensure that the object is a cow.
Clarity and Identification: Only annotate if the object clearly resembles a cow. Do not include any objects that might be confused with other elements like rocks or terrain features unless you can clearly distinguish the cow-like form.
While annotating, ensure each cow is bounded separately, avoiding overlaps in annotation with other cows or labels.