HiDream AI's image-to-image editing model for instruction-based image modifications and transformations
Settled by community votes across 4 shared challenges, with an AI judge weighing in on each.
HiDream E1
#33 of 32 in Image Editing
Vidu Q2
#42 of 62 in Text-to-Image
Where the votes landed
HiDream E1
0%
win rate
Ties
0%
Vidu Q2
0%
win rate
Challenge by challenge
The strongest take from each model on every shared challenge, with the AI judge's read.
Bald man challenge
Image Editing“Give the person a full, thick head of natural hair with realistic texture, density, and a natural hairline. Preserve facial features and lighting.”
AI Judge Analysis
HiDream E1
- + Successfully added a large volume of hair.
- − Failed to preserve the source image's background, changing it to a desert landscape.
- − The skin texture on the forehead is heavily distorted with unnatural artifacts.
- − The hair texture looks AI-smoothed and lacks realistic integration with the scalp.
Vidu Q2
- + Excellent preservation of the source background, jacket, and facial details.
- + The added hair has a realistic texture and a very natural-looking hairline.
- + Lighting on the new hair perfectly matches the existing environment and beard.
- − The transition area above the right ear (viewer's left) is slightly abrupt.
Verdict: Vidu Q2 is the clear winner as it successfully applied the requested edit while perfectly preserving the original background and facial features of the source image. In contrast, HiDream E1 completely changed the background environment and introduced significant artifacts to the subject's face, failing the 'preserve source' requirement.
Over-the-top cartoon caricature
Editing“Create a caricature of me and my job. Make it exaggerated and humorous, incorporating my profession as a tv show anchor and my love for dogs and hockey.”
AI Judge Analysis
HiDream E1
- + Successfully incorporates all elements: hockey setting, dog, and news anchor blazer.
- + High level of humorous exaggeration in facial features and hair.
- + Captures the 'TV anchor' vibe with the lapel mic and professional attire.
- − The facial resemblance to the source image is significantly diminished.
- − Minor anatomical issues with the hands/fingers holding the dog.
Vidu Q2
- + Preserves the subject's facial features much more accurately than the competitor.
- + Cleverly includes small details like paw prints on the news script and a puck on the desk.
- + Excellent artistic style that feels like a professional digital caricature.
- − The hand holding the microphone has six fingers.
- − Subject's clothing remains a casual denim shirt rather than the typical anchor attire.
Verdict: Vidu Q2 is the winner because it maintains a much better likeness of the woman in the source image while delivering a high-quality, creative caricature. While HiDream E1 did a better job with the news anchor attire, Vidu Q2 followed the hockey and dog themes more inventively with details throughout the scene despite a major anatomical error on the hand.
Studio Ghibli Anime Style
Editing“Transform this photo into a Studio Ghibli–inspired illustration. Use soft pastel colors, hand-painted textures, gentle lighting, dreamy backgrounds, and a warm, nostalgic mood”
AI Judge Analysis
HiDream E1
- + Successfully captures the pastel color palette requested.
- + Maintains the composition and recognizable poses of the original meme.
- − The art style is more generic modern webcomic/pop-art than Studio Ghibli.
- − Loses significant detail in the clothing and character expressions.
- − Changes the colors of the clothing unnecessarily.
Vidu Q2
- + Executes a much more authentic hand-painted texture reminiscent of Ghibli watercolors.
- + Perfectly preserves the specific clothing details, including the complex plaid pattern of the shirt.
- + Maintains the exact facial expressions and lighting from the source image while stylizing them.
- − The background is slightly less 'dreamy' compared to typical Ghibli film stills.
Verdict: Vidu Q2 is the clear winner as it successfully applies the requested Studio Ghibli art style with beautiful hand-painted textures while perfectly preserving every detail of the original meme, including the plaid shirt pattern. HiDream E1 adopts a more generic cartoon style that loses much of the source image's character and changes colors for no clear reason.
Golden Hour Stroll
Image Editing“Add dynamic motion to this photo: make hair blow in the wind, add leaves flying, energetic and lively feel.”
AI Judge Analysis
HiDream E1
- + Successfully applied the 'hair blowing in wind' effect with strong visual impact.
- + The overall image quality remains sharp.
- − The added leaves are stylistically inconsistent with the photo, looking more like vector clipart than real leaves.
- − Significant facial distortion occurred, making the subject look like a different person than in the source image.
- − The image resolution and aspect ratio were altered slightly.
Vidu Q2
- + Excellent source preservation; the woman and dog are identical to the original image.
- + The added leaves look like natural, photographic elements that blend well with the scene.
- + Effective hair motion that feels realistic rather than stylized.
- − The quantity of leaves is perhaps slightly high, which can feel a bit cluttered.
Verdict: Vidu Q2 is the clear winner as it successfully implemented all parts of the edit prompt while preserving the identity of the person and the quality of the original photo perfectly. HiDream E1 failed the preservation check by significantly altering the woman's face and adding leaves that look like digital illustrations instead of photographic elements.
Explore each model
ShengShu Technology's text-to-image and reference-to-image model with support for character consistency and multi-reference image processing