Are these expressions melded manually? I'm interested in knowing what it means for a simulation program to model the mixing of anger and sadness, for example.
keeratsingh2002
In implementing blend shapes for animated characters, how do animators decide on the number and variety of key shapes necessary to convey a wide range of emotions and expressions effectively?
weinatalie
How do animators combat the uncanny valley effect, especially when conveying something as complex as human facial expressions? I would imagine that—especially when using a technique like blend shapes—it would be easy to fall into the trap of creating an expression that looks just a bit off. An animator would probably have to go back in and touch up the result, especially since human expressions are not really perfectly interpolated in practice.
everbolt
Is there a standardized set of areas on the face that animators focus on when depicting a facial expression? I'd imagine for human faces, there are common areas (forehead, cheeks) that often change for any emotion. How does this translate to animating facial expressions for non-human entities? Animation instances like Toy Story seem a lot less trivial when animating natural-looking facial expressions of unnatural entities (dogs, alien toys, etc).
yykkcc
I'm quite astonished by the versatility and depth that blend shapes bring to character animation. It's intriguing to see how complex expressions can be crafted by combining fundamental emotions. For instance, I hadn't considered how mixing an expression of anger with surprise could articulate a 'What the...?!' reaction. Based on this feature, can we generalize this interpolation to other areas?
GarciaEricS
This reminds me about the concept of an "eigenface" in computer vision. It is essentially one of the eigenvevtors when you run PCA on a dataset of faces. Those faces encode information about the most common face patterns. You combine faces like you do here with linear combinations to arrive at new faces. Interesting stuff
rohan19a
I wonder if we could create some new never before seen emotions by combining faces together!
Are these expressions melded manually? I'm interested in knowing what it means for a simulation program to model the mixing of anger and sadness, for example.
In implementing blend shapes for animated characters, how do animators decide on the number and variety of key shapes necessary to convey a wide range of emotions and expressions effectively?
How do animators combat the uncanny valley effect, especially when conveying something as complex as human facial expressions? I would imagine that—especially when using a technique like blend shapes—it would be easy to fall into the trap of creating an expression that looks just a bit off. An animator would probably have to go back in and touch up the result, especially since human expressions are not really perfectly interpolated in practice.
Is there a standardized set of areas on the face that animators focus on when depicting a facial expression? I'd imagine for human faces, there are common areas (forehead, cheeks) that often change for any emotion. How does this translate to animating facial expressions for non-human entities? Animation instances like Toy Story seem a lot less trivial when animating natural-looking facial expressions of unnatural entities (dogs, alien toys, etc).
I'm quite astonished by the versatility and depth that blend shapes bring to character animation. It's intriguing to see how complex expressions can be crafted by combining fundamental emotions. For instance, I hadn't considered how mixing an expression of anger with surprise could articulate a 'What the...?!' reaction. Based on this feature, can we generalize this interpolation to other areas?
This reminds me about the concept of an "eigenface" in computer vision. It is essentially one of the eigenvevtors when you run PCA on a dataset of faces. Those faces encode information about the most common face patterns. You combine faces like you do here with linear combinations to arrive at new faces. Interesting stuff
I wonder if we could create some new never before seen emotions by combining faces together!