Salient Imagenet

Class: 240 (Appenzeller Sennenhund), Feature: 1023

This feature is core for class Appenzeller Sennenhund. It is the most important feature for class Appenzeller Sennenhund.

Below, we display the five Appenzeller Sennenhund images that most highly activate the feature, along with the Neural Activation Map (heatmap) and Feature Attack for each image. These are the same as the images shown to crowdworkers to annotate this class-feature pair.

We additionally provide the full responses of the five crowdworkers that annotated this pair, as well as links to other related pairs (i.e. same class or same feature).

MTurk Worker Annotations

core

spurious

core

spurious

Main Focus	Confidence	Explanation
Main Object	5	The focus was on the face of the dog, the main object.
Main Object	4	Focus is on nose of a dog, focus is on behind of a dog
Main Object	4	Highlighted visual attributes is on main object that is dogs face, tail and feet
Main Object	4	Focus is on the left side of the eyes and the part just below from where the tail of the appenzeller starts.
Main Object	2	The focus is on the tail which is part of the appenzeller

core

Related Class-Feature Pairs

Important Features for Appenzeller Sennenhund

There are 5 features annotated as core and 0 features annotated as spurious for Appenzeller Sennenhund. Core features include 1023, 1287, 1313, 1385, 2003.

Classes where Feature 1023 is among Most Important (i.e. top-5)

Feature 1023 is core for the following classes: Appenzeller Sennenhund, Entlebucher Sennenhund, Basenji.