Detecting Avocados to Zucchinis: What Have We Done, and Where Are We Going?
Citations
30,811 citations
30,462 citations
Cites methods from "Detecting Avocados to Zucchinis: Wh..."
...Unlike previous datasets containing entry-level categories [29], such as “dog” or “chair,” like [28], ImageNet used the WordNet Hierarchy [30] to obtain both entry-level and fine-grained [31] categories....
[...]
...ImageNet was created to capture a large number of object categories, many of which are fine-grained....
[...]
...For instance the ImageNet dataset [1], which contains an unprecedented number of images, has recently enabled breakthroughs in both object classification and detection research [5], [6], [7]....
[...]
...Recently, a detection challenge has been created from 200 object categories using a subset of 400,000 images from ImageNet [34]....
[...]
...MS COCO contains considerably more object instances per image (7.7) as compared to ImageNet (3.0) and PASCAL (2.3)....
[...]
1,109 citations
691 citations
519 citations
Cites background or methods from "Detecting Avocados to Zucchinis: Wh..."
...Appendix B and (Russakovsky et al., 2013) have additional comparisons....
[...]
...The x-axis corresponds to object properties annotated by human labelers for each object class (Russakovsky et al. 2013) and illustrated in Fig....
[...]
...Human subjects annotated each of the 1000 image classification and single-object localization object classes from ILSVRC2012-2014with these properties (Russakovsky et al. 2013)....
[...]
...The other properties were computed by asking human subjects to annotate each of the 1000 object categories (Russakovsky et al. 2013)...
[...]
References
73,978 citations
49,639 citations
46,906 citations
15,935 citations
"Detecting Avocados to Zucchinis: Wh..." refers background in this paper
...Others have provided insight into dataset design [14, 18, 7]....
[...]
...The field of large-scale categorical object detection is rapidly growing [7, 3, 12], both by developing more robust object representations and also by collecting richer datasets for training and evaluation....
[...]
...The most in-depth analysis of generic object detection to date has been performed on the PASCAL challenge in [7] and especially in [10]....
[...]
...The commonly accepted measure of detection accuracy of an algorithm consists of two requirements [7]: (1) all instances of an object class should be correctly localized, where an object instance with a bounding box B̂ is considered correctly localized if a window B returned by the algorithm satisfies the intersection over union (IOU) measure: IOU(B̂, B) = area(B̂∩B) area(B̂∪B) ≥ 0....
[...]
..., PASCAL [7], LabelMe [15], TinyImages [19], SUN [24], and ImageNet [4]....
[...]
14,708 citations
"Detecting Avocados to Zucchinis: Wh..." refers methods in this paper
...It uses an image classification system with dense SIFT features and color statistics [13], a Fisher vector representation [16], and a linear SVM classifier, plus additional insights from [2, 17], combined with the deformable parts-based model (DPM) [9] which has been the dominant model for generic object detection for many years....
[...]