Large-Scale Image Retrieval with Attentive Deep Local Features
Citations
611 citations
Cites methods from "Large-Scale Image Retrieval with At..."
...[47] to learn global image descriptors using a saliency mask....
[...]
607 citations
594 citations
Cites background or methods or result from "Large-Scale Image Retrieval with At..."
...Several of these methods start by dense descriptor extraction [3,37,60,61] and later aggregate these descriptors into a compact image-level descriptor for retrieval....
[...]
...Works most related to our approach are [37,60]: [37] develops an approach similar to ours, where an attention module is added on top of the dense description stage to perform keypoint selection....
[...]
...We compare against upright RootSIFT descriptors extracted from DoG keypoints [29], HardNet++ descriptors with HesAffNet features [34, 35], DELF [37], SuperPoint [13] and DenseSfM [45]....
[...]
...As in other previous work [37,43,58], the most straightforward interpretation of the 3D tensor F is as a dense set of descriptor vectors d:...
[...]
...DELF does not refine its keypoint positions - thus, detecting the same pixel positions at feature map level yields perfect accuracy for strict thresholds....
[...]
378 citations
Cites methods from "Large-Scale Image Retrieval with At..."
...The images from both Google Landmarks [36] and Berkeley Deep Drive [58] are resized to 640×480 and converted to grayscale....
[...]
...We thus train on 185k images from the Google Landmarks dataset [36], containing a wide variety of day-time urban scenes, and 37k images from the night and dawn sequences of the Berkeley Deep Drive dataset [58], composed of road scenes with motion blur....
[...]
...The images from both Google Landmarks [11] and Berkeley Deep Drive [21] are resized to 640×480 and converted to grayscale....
[...]
330 citations
Cites background from "Large-Scale Image Retrieval with At..."
...Already in 2D, learned descriptors significantly outperform their engineered counterparts [49, 28]....
[...]
References
123,388 citations
49,914 citations
46,906 citations
30,811 citations
23,396 citations
"Large-Scale Image Retrieval with At..." refers methods in this paper
...Finally, we perform geometric verification using RANSAC [10] and employ the number of inliers as the score for retrieved images....
[...]